- Email: [email protected]
Polymer-modified microfluidic immunochip for enhanced electrochemical detection of troponin I
Sensors and Actuators B 209 (2015) 478–485
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Polymer-modified microfluidic immunochip for enhanced electrochemical detection of troponin I Josef Horak a,∗ , Can Dincer a,b , Edvina Qelibari a , Hüseyin Bakirci a , Gerald Urban a,b a b
Laboratory for Sensors, Department of Microsystems Engineering (IMTEK), University of Freiburg, Georges-Köhler-Allee 103, 79110 Freiburg, Germany Freiburg Materials Research Center (FMF), University of Freiburg, Stefan-Meier-Straße 21, 79104 Freiburg, Germany
a r t i c l e
i n f o
Article history: Received 16 October 2014 Received in revised form 29 November 2014 Accepted 2 December 2014 Available online 11 December 2014 Keywords: Immunochip Immunoassay Electrochemical detection Polyethylenimine Troponin I
a b s t r a c t We report enhanced electrochemical detection of a cardiac biomarker using a combination of a disposable microfluidic immunochip fabricated in Vacrel® 8100 photoresist film and a highly effective surface functionalization employing polyethylenimine (PEI). The use of the photoresist film enables fast prototyping and low-cost production without the need for a cleanroom. We use the surface carboxylates of the photoresist to biofunctionalize the microchannel on the chip using direct amine-specific coupling and modification via adsorbed and immobilized PEI in both linear (LPEI) and branched (BPEI) form. Characteristics of each immobilization strategy are assessed by a sandwich immunoassay for troponin I quantification in serum. The best assay performance is achieved using the immunochip modified with immobilized BPEI: the antigen was routinely detected at concentrations of 25 pg ml−1 in 4 min read-out time and 5 l serum sample, representing an 18-fold improvement of the detection limit and 2.5-times faster read-out time in comparison to the assay implemented over amine-reactive esters without the PEI coating. We demonstrate that immobilized BPEI represents a stable and tunable scaffold that enhances biomolecule immobilization up to 60-times and stability 2-fold while lowering non-specific binding by factor of three. This system provides a versatile and nonaggressive means for incorporation of biological material to virtually any platform design. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The development of a versatile immunosensing platform that can effectively and economically detect a broad range of different molecular targets is highly desirable. Performance characteristics of the vast majority of immunosensors, and biosensors in general, are directly linked to the choice of materials and applicable immobilization methods to attach the immunoreactants onto solid surfaces. The ideal immobilization strategy should employ mild chemical conditions and allow large quantities of biomolecules to be immobilized while retaining their biological activity. The immobilization surface should provide a large area for antibody–antigen interaction within a small total volume and should limit the non-specific protein adsorption. In terms of practical use, immobilization methods should extend the shelf life, which is a necessary criterion for commercial biosensors such as point-of-care (POC) diagnostics. Without question, polystyrene, silica and polysaccharides are the three most common substrates for biomolecule
∗ Corresponding author. Tel.: +49 0761 203 7264; fax: +49 0761 203 7262. E-mail address: [email protected] (J. Horak). http://dx.doi.org/10.1016/j.snb.2014.12.006 0925-4005/© 2014 Elsevier B.V. All rights reserved.
attachment. Polystyrene is the traditional substrate for enzymelinked immunosorbent assays (ELISA) in the form of microtiter plates or beads [1,2], but has been employed in biosensors as well [3]. Silica can be fabricated into a number of different forms including particles, wafers and thin films deposited by plasma chemical vapor deposition. Silica particles are employed in both chromatography and immunoassays [4,5], while silica wafers and thin films are increasingly being used in biosensors [6,7]. The same issues arise in both traditional immunoassays and biosensors: the decreased surface activity of the immobilized antibodies (or biomolecules in general) and non-specific binding (NSB) of various sample components to the solid phase. Antibodies attached by passive adsorption to polystyrene undergo conformational changes and remain only marginally active (<10%) [8,9]. The covalent immobilization used for silica and for PDMS relies predominantly on silanization protocols [10], which is a time consuming procedure and requires aggressive conditions that are not always compatible with non-silica parts of the packaging or housing [11]. These can have a negative impact on both the assay sensitivity and NSB – the former because relatively few active capture molecules are available to bind the analyte, and the latter because sample components are inclined to interact nonspecifically with the
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
preponderance of inactive antibodies, which may be partially denatured. To tackle these problems, various polymers have been used for biosensor applications. Instead of attaching proteins onto a twodimensional solid surface, biomolecules can diffuse into a porous matrix formed by polymer membranes or surface coatings [12]. Polysaccharides were employed in a number of different types of chromatographic supports and immobilization matrices in biosensors such as carboxymethyl dextran in Biacore systems [13,14]. The traditional membrane material for the vast majority of lateral flow assays is nitrocellulose; nylon and polyvinylidene fluoride membranes had only limited success due to various factors such as high cost and limited utility [12,15]. Based on the aforementioned considerations, we merged the above approaches into a single platform to simultaneously take advantage of our microfluidic immunochip [16] and a novel immobilization strategy. We combined an microfabrication strategy using the flexible dry film photoresist, Vacrel 8100® , which offers good wetting properties for passive fluid control and easy biofunctionalization, together with an immobilization strategy in which the surface is first coated with a thin film of polyethylenimine that acts as a matrix for immobilization of biomolecules. We will refer to such a polymer coating as a “passivating” layer since it limits the interaction of proteins and other sample components with the surface – a fitting term used by Herron at al. [9]. Polyethylenimines are polycationic aliphatic polymers that can be produced either in linear form (LPEI) or with varying degrees of branching (BPEI). PEI has a high density of amino groups – every third atom in the polymer chain is a nitrogen atom. BPEI contains primary, secondary and tertiary amino groups with a ratio of 1:2:1, while LPEI mostly contains only secondary amines. In general, PEIs are both highly basic and water soluble polymers available in molecular weights ranging from 700 Da to 1000 kDa. They have been extensively studied predominantly in medicinal chemistry [17], but also for non-pharmaceutical use [18,19]. Importantly for our application, the polymer has been shown to be effective for immobilization of biocatalysts to solid supports by adsorption [20,21] and to have a positive effect on enzyme activity and stability [22] while having unique protein-resistant properties [23]. In our previous report, we demonstrated the high level of performance of our immunosensing platform using a competitive ELISA for substance P (an analyte of low molecular weight) and simple carbodiimide-based coupling chemistry [24]. To assess the benefit of the PEI modified immunosensor, we present here rapid detection with clinically relevant sensitivity of troponin I (cTnI) in a sandwich assay format. cTnI is the “gold standard” for diagnosis of cardiac muscle cell damage and death [25]. cTnI (23.8 kDa) is a component of the troponin ITC complex (77.2 kDa), further comprising troponin T (35 kDa) and troponin C (18.4 kDa), a heteromeric protein that plays an important role in the regulation of skeletal and cardiac muscle contraction [26,27]. cTnI is expressed only in myocardium and there are no examples known of cTnI expression in healthy or injured skeletal muscle or in other tissue types [28,29], hence it is currently widely used for the diagnosis of acute myocardial infarction, unstable angina, post-surgery myocardium trauma, as well as several other diseases related to cardiac muscle injury [30].
2. Material and methods
479
from HyTest Ltd., Finland. Serum and blood samples were obtained from University Medical Center Freiburg. Glucose oxidase avidin conjugate (GOx-avidin) was obtained from Biomol, Germany. All other chemicals were purchased from Sigma–Aldrich or otherwise as stated in the text. 2.2. The chip fabrication A detailed description of the wafer fabrication process utilizing Vacrel® 8100 can be found in our previous report [16]. 2.3. Microtiter plate-based troponin I ELISA See S1, Supplementary material for further information. Every step of the immunoreactions was followed by three washing cycles on a commercial ELISA microplate washer using wash buffer (10 mM PBS, 138 mM NaCl, 2.7 mM KCl, 0.005% Tween 20, pH 7.4). Standard coating volume of 100 l per well was used for all immunoreactants, 200 l was used for blocking. Plates (NuncImmunoTM LockWellTM Modules, Nunc GmbH & Co. KG, Germany) were incubated sealed on a plate shaker preheated to 25 ◦ C and at 450 rpm. Immunoreactants were diluted in PBS buffer (10 mM PBS, 138 mM NaCl, 2.7 mM KCl, pH 7.4). Microtiter plate was coated with capture antibody combination, 16A11 and 19C7, to a final concentration of 10 g ml−1 diluted in PBS. After 2 h incubation, the plate was washed. To saturate any binding sites not already occupied by the capture antibody, incubation with 200 l of blocking solution was carried out for 30 min at 25 ◦ C on a microplate shaker followed by overnight incubation at 4 ◦ C without shaking. After subsequent washing, the plate was ready to use. Troponin ITC complex was spiked into cTnI-free serum to concentration range 0.02–12.5 ng ml−1 using a 5-fold serial dilution. The calibrators were incubated for 2 h followed by a washing step and 1 h incubation with 4 g ml−1 of biotinylated detection antibody, MF4b. 100 l of a solution of GOx conjugated avidin diluted to 1 g ml−1 in reagent diluent was added and the plates were incubated for 30 min at 25 ◦ C. After subsequent washing, 50 l of ABTS diluted in phosphate-citrate buffer (200 mM Na2 HPO4 , 100 mM citric acid, pH 5.6) to concentration 2.2 mg ml−1 were pipetted into each well. Solutions containing 160 mM of glucose and 0.8 mg ml−1 of horseradish peroxidase both in phosphatecitrate buffer were mixed in 1:1 ratio. A volume of 50 l of this solution was pipetted into each well. The plate was incubated without shaking for 1 h at 25 ◦ C and then immediately measured on an absorbance microplate reader (SpectraMax 340PC384, Molecular Devices GmbH, Germany) preheated at 25 ◦ C at 405 nm. Software SoftMax Pro (Molecular Devices GmbH, Germany) was used to collect the data. 2.4. Chip-based troponin I ELISA For every incubation step, a 5 l of (immuno)reagents was pipetted on the chip inlet to fill the chip immobilization capillary. Chips were incubated in closed Petri dishes at 25 ◦ C. Every step of the immunoreactions was followed by a washing step with a custom made vacuum pump using 200 l of a wash buffer. PBS buffer and wash buffer were identical to those used for microtiter plate ELISA. Methods for all immobilization strategies tested are described in S2, Supplementary material.
2.1. Materials and chemicals Pyralux® AP, Vacrel® 8100 and Teflon® 1600 AF were purchased from DuPontTM . Monoclonal capture and detection antibodies (mAb); troponin ITC complex and cTnI-free serum were purchased
2.4.1. Direct immobilization strategy The immobilization area of the chip was pretreated for 10 min with 10% Na2 CO3 and washed with 100 mM 2-(Nmorpholino)ethanesulfonic acid (MES) buffer, pH 6.0. The
480
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
capillary was treated with activation buffer (100 mM 1-ethyl-3-[3dimethylaminopropyl]carbodiimide hydrochloride (EDC), 200 mM N-hydroxysulfosuccinimide (SNHS) solution buffered at pH 6.0 with 100 mM MES, 0.9% NaCl) for 60 min. A 50 g ml−1 solution of 19C7 and 16A11 mAbs were prepared in PBS buffer and incubated for 60 min in the activated capillary. After a washing step, the capillary was blocked for 30 min using 1% BSA in PBS buffer. The rest of the assay workflow was identical with the conventional cTnI ELISA described above, except for the higher utilized concentration of the GOx-avidin label (4 g ml−1 ) and duration of incubation steps cut by half. 2.4.2. Passivating layer method utilizing immobilized BPEI The immobilization area of the chip was pretreated for 10 min with 10% Na2 CO3 and washed with 100 mM MES buffer, pH 6.0. The capillary was treated with activation buffer (100 mM EDC, 200 mM sulfo-NHS solution buffered at pH 6.0 with 100 mM MES, 0.9% NaCl) for 60 min. 5% (w/v) solution of BPEI was prepared in PBS buffer, vortexed well and neutralized with 6 M HCl to pH 8. After 60 min incubation at 25 ◦ C in the activated capillary, the channel was washed with 200 l of 1 M NaCl in ultra-pure H2 O and 200 l of PBS. The BPEI was reacted with 2.5% glutaraldehyde (GA) solution containing 10 mM NaCNBH3 for 1 h and washed with 200 l of ultra-pure H2 O and 100 l of 50 mM carbonate–bicarbonate buffer, pH 9.6. A 5 g ml−1 solution of 19C7 and 16A11 mAbs were prepared in 50 mM carbonate–bicarbonate buffer, pH 9.6 and incubated in the activated channel for 60 min. Chips were washed with 200 l of 50 mM carbonate–bicarbonate pH 9.6 and incubated for 30 min with 25 mM NaCNBH3 prepared in PBS buffer. After washing with 100 l of PBS, chips were blocked with 1% BSA blocking solution containing 25 mM of NaCNBH3 for 30 min followed by washing. The rest of the assay workflow was identical with the conventional cTnI ELISA described above, except for the duration of incubation steps cut by half. 2.4.3. Amperometric measurement Chips were fastened into a custom-made adapter connected to a potentiostat (Jobst Technologies, Germany) and a syringe pump (PHD2000 Harvard Aparatus; syringes Hamilton Company); constant flow (withdraw mode) of 15 l min−1 was used throughout the measurements. 100 mM PBS, 100 mM NaCl, pH 7.4 buffer was used to establish the baseline, glucose solution substrate (40 mM glucose in 100 mM PBS, 100 mM NaCl, pH 7.4) was used to generate the electrochemical signal. The measurement was recorded with bioMON software (Jobst Technologies, Germany) at +350 mV vs. Ag/AgCl on-chip pseudo-reference electrode. 3. Results and discussion In our previous reports, we have shown the high degree of performance of our immunosensing platform using qualitative direct ELISA for Epstein-Bar virus detection [16] and quantitative competitive ELISA for small peptide substance P [23]. To complete the demonstration that the immunochip is a versatile platform applicable to all of the major analyte classes, we investigated whether it could detect protein biomarkers, using troponin I as a model system. 3.1. Immunochip design and fabrication As can be seen in Fig. 1a, a single chip consists of a microchannel, an electrochemical measurement cell containing Pt working, counter and Ag/AgCl pseudo-reference electrodes (Fig. 1b), and
contact pads. The immobilization area of the channel (Fig. 1c), bordered by the inlet port and the stopping barrier, possesses a surface of 112 mm2 and 3.6 l volume, which creates a high surface to volume ratio of 310 cm−1 . Detailed geometry and dimensions of the electrodes and microfluidic channel are depicted in Fig. S3. For the bioassay preparation, all reagents are pipetted onto the inlet and enter the immobilization channel by capillary force, so the reagents coat a well-defined area and do not contaminate the electrochemical measurement cell. The flexible microfluidic chip is produced completely on waferlevel by a standard photolithographic technique employing dry film resist Vacrel® 8100. Briefly, this fabrication strategy relies on lamination of the film photoresist to a platinum patterned polyimide wafer using heat and pressure. All chip structures are subsequently generated by photolithography. A single 6-inch wafer comprises 40 chips. The most apparent advantages of using dry film resist are fast processing and prototyping requiring minimal material and device costs; except for Pt deposition, all fabrication steps were performed outside of a cleanroom. The less apparent advantage of Vacrel® 8100 is the possibility of using its surface carboxylates for amine-specific “direct” biofunctionalization. 3.2. Development of troponin I ELISA The crucial issue of the cTnI assay performance is the choice of an antibody combination that would recognize the cTnI epitopes that are the least affected by (i) conformational changes and (ii) the complexation of the antigen when released into the patient’s blood from necrotic tissue after acute myocardial infarction [31]. The developed cTnI ELISA is based on the standard sandwich assay format. Two mouse monoclonal antibodies (mAb), 19C7 and 16A11, binding epitopes located in the central stable part of the cTnI molecule (between 30 and 110 amino acid residues) were used for solid phase immobilization. A third high affinity mAb conjugated biotin, MF4b (recognizing epitope 190–196), was used for detection. In the last step, the sandwich complex was labeled with a GOx conjugated avidin. The subsequent implementation of the cTnI ELISA on the chip and using electrochemical detection were ample reasons to develop the assay with GOx instead of the more common horseradish peroxidase label. The assay concept and antibody combination were first optimized on a microtiter plate before being implemented on the chip (S1, Supplementary material). Since the cTnI assays rely on serum troponin level [32,33], undiluted cTnI-free serum was used to evaluate the assay parameter both on the plate and the immunochip. However, measurements performed with spiked whole blood and serum samples showed statistically indistinguishable values of cTnI in the entire calibration range (Fig. S2). International guidelines endorsed total cTnI assay imprecision at their 99th percentile reference value to be ≤10% coefficient of variation (CV) [34]. In our setting, the assessment of the 99th percentile was not possible (ranges between 0.01 and 0.3 ng ml−1 for commercial assays) [35,36]. Our assay performed with 11.6% CV at 0.02 ng ml−1 in serum; diagnostic levels of cTnI are considered negative up to 0.1 ng ml−1 . 3.3. On-chip implementation of troponin I ELISA We experimented with five different strategies to immobilize the biological probes on Vacrel® 8100 which composes the immobilization area of the immunochip capillary. All methods employed random covalent coupling via primary amines present predominantly on lysine residues of the biomolecules; these included: (i) direct covalent coupling to the photoresist material via carboxyl-to-amine crosslinking, (ii) covalent attachment to adsorbed LPEI, (iii) covalent attachment to immobilized LPEI, (iv) covalent attachment to adsorbed BPEI and
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
481
Fig. 1. Summary of the cTnI ELISA implementation on the BPEI modified immunochip. (a) Photograph of the utilized immunochip (image without front cover) showing the main dimensions and functional elements. (b) The electrochemical measurement cell containing platinum working (WE), counter (CE) and Ag/AgCl reference (RE) electrodes. (c) The microfluidic channel. The surface modification and assay preparation takes place in the immobilization capillary (highlighted gray) bordered by the inlet and stopping barrier. (d) Direct immobilization strategy and (e) passivating layer method utilizing immobilized BPEI are shown.
Fig. 2. Comparison of immobilization approaches tested by performing on-chip cTnI assay. Signals represent 4 min stop-flow. Signal-to-background ratio (SBR) was calculated for measurements where positive signals were higher than a control sample. Error bars represent SD and are collected from three parallel measurements.
finally (v) covalent attachment to immobilized BPEI. Alongside the cTnI assay, the immobilization methods were evaluated using GOx as a model biomolecule. This enzyme is well characterized and commonly used for biosensing applications including ours [37]. Further text will be restricted only to the comparison of the direct immobilization strategy (Fig. 1d) with the passivating layer method utilizing covalently attached BPEI (Fig. 1e). All other approaches (S2, Supplementary material) failed to provide satisfactory results (Fig. 2) except for covalent protein attachment to adsorbed LPEI. The reason for this was the generally lower stability of the adsorbed PEIs to the high number of washing steps required for immunoassay preparation and in the long-term, since the PEI-Vacrel® interactions are predominantly electrostatic in nature. As shown in S4, Supplementary material, the immobilized BPEI was stable for six days while adsorbed LPEI only for two days. Nevertheless, the immobilization strategy via adsorbed LPEI showed potential for the efficient attachment of biomolecules in applications that do not involve such a high number of washing steps.
3.3.1. Direct immobilization method Briefly, the utilized photoresist formulation contains acrylic binders and various multifunctional acrylates possessing carboxylic and carboxylate groups. These can be reacted with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and Nhydroxysulfosuccinimide to form an reactive O-acylisourea and succinimidyl ester intermediates that are easily displaceable by nucleophilic attack from primary amines of the target protein. SNHS ester is considered to be more stable than the O-acylisourea intermediate while allowing for efficient conjugation to primary amines at physiologic pH. The direct immobilization strategy represents a fast and convenient way of biomolecule attachment, nevertheless, requires a higher concentration of the probe. The optimized on-chip cTnI assay utilized higher coating concentration of the capture mAbs combination (50 g ml−1 ) and the GOx-avidin label (4 g ml−1 ) in comparison to the microplate assay. BSA was used to simultaneously block the unreacted crosslinker intermediates and as a subsequent blocking step. Storage of such pre-immobilized and pre-blocked chips strongly depended on the stability of the immobilized biomolecules. Stored at 4 ◦ C, blocked chips containing 19C7/16A11 mAbs had to be used within 24 h; in the case of GOx, chips had to be used immediately upon preparation.
3.3.2. Passivating layer method using immobilized BPEI In this immobilization strategy, the cTnI specific capture mAbs were coupled to the capillary surface coated with BPEI (25,000 Mw). Our choice was motivated by several factors: (i) PEIs of higher molecular weight are relatively inert to protein adsorption [23] and for some proteins, e.g. glucose oxidase, can resist the non-specific adsorption even without a dedicated blocking strategy (S4, Supplementary material). (ii) PEIs can be readily chemically functionalized with proteins of interest. (iii) Both the protein blocking properties and functionalization efficiency can be optimized by the PEI concentration utilized. The primary and secondary amino groups of BPEI were used both for immobilizing the polymer to the EDC/SNHS activated channel surface and for conjugating the biomolecules to the BPEI via glutaraldehyde.
482
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485 Table 1 cTnI assay parameters for chip-based direct and passivating immobilization strategies compared to the conventional microtiter plate. Parameters
Chip (direct)
Chip (passivating)
Microtiter plate
LOD (ng ml−1 ) LOQ (ng ml−1 ) Max. imprecision (%CV) Imprecision at LOD (%CV) IC50 (ng ml−1 ) Slope (nA ng ml−1 ) Total assay time (h) Reagent/sample volume (l)a
0.45 0.75 ≤20 ≤14 85.4 0.62 5
0.025 0.042 ≤20 ≤12 6.36 1.01 6
0.019 0.03 ≤12 ≤12 20.62 1.21 6 + overnightb 100/200c
a b c
Fig. 3. Comparison of immobilization/surface activity and stability of biomolecules (GOx) coupled on the immunochip via direct and BPEI immobilization strategies. Signals represent 5 min stop-flow. Error bars represent SD and are collected from three parallel measurements.
Prior to the BPEI immobilization, the basic BPEI solution (pH 10.5) had to be neutralized to pH 8 in order to prolong the stability of the succinimidyl ester, which is the amine-reactive intermediate of the activated surface carboxylates. The half-life of the succinimidyl ester at this pH value is approximately 1 h, providing enough time for successful BPEI immobilization. However, already at pH 8.6, the stability of SNHS ester decreases to only 10 min [38]. After the immobilization step, wash solution containing 1 M sodium chloride was used to remove the unbound BPEI from the immobilization capillary surface. The biofunctionalization of BPEI had to be optimized for different protein targets. The immobilization of antibodies on GA activated BPEI required utilization of immobilization buffer at pH 9.6 to deprotonate the lysine -amino group (pKa 10.5 in polypeptides). Moreover, we used an additional blocking step, simultaneously deactivating the unreacted GA aldehyde groups with BSA. However, for GOx conjugation, only a negligible difference in immobilization efficiency was observed for different buffer pH (pH 7.4 vs. 9.6). Although formation of an unstable Schiff base on both ends of the monomeric GA has been ruled out as a mechanism for GA crosslinking with proteins [39], it was necessary to use sodium cyanoborhydride to convert the Schiff base to a stable secondary amine. Use of the BPEI passivating layer enabled lowering the concentration of the immobilized capture antibodies (5 g ml−1 ) and prolonging the stability of the immobilized biomolecules. In comparison to the direct immobilization method, the use of a passivating layer extended the storage of pre-immobilized and preblocked chips to 48 h at 4 ◦ C. In the case of GOx, as can be seen in Fig. 3, the effect of PEI on long-term stability was significant; nevertheless, immobilized GOx lost about 64% of its activity (93% with direct immobilization) in 24 h when stored at 4 ◦ C. As mentioned earlier, covalent attachment of the PEI to the capillary surface prolongs its stability and usability to six days at 4 ◦ C. The above mentioned effect on prolonged biomolecule stability was significant and may be beneficial for a variety of applications performed in laboratory conditions where high biomolecule loading is of interest. Nevertheless, this immobilization strategy is not sufficiently stable for long-term storage, as may be required for commercial POC devices, such as lateral flow immunoassay systems using predominantly nitrocellulose as a protein immobilization matrix. Specifically, various qualitative tests for troponin T and troponin I in the form of sticks or cassettes are marketed with a
5
Volume per well/chip. Blocking step. Volume used for plate blocking.
declared stability of several weeks at ambient temperature. On the other hand, these traditionally designed lateral flow tests suffer from performance limitations; most notably sensitivity and reproducibility limits their applications in quantitative systems [15]. 3.4. The electrochemical immunoassay of troponin I To determine the effect of the immobilization strategies, we have established the assay parameters by performing cTnI calibration curves in undiluted human serum. Mimicking the physiological conditions, cTnI, in the form of human cardiac troponin ITC complex, was spiked into undiluted human serum over the concentration range 20 pg ml−1 to 12.5 ng ml−1 using a 5-fold serial dilution and pipetted on the preimmobilized chips as described in Section 2. After a washing step, MF4-biotin detection antibody and GOx-avidin were subsequently added, followed by a final washing. For all immunoreactions, we used 5 l of reagent volume and incubation times were cut to half in comparison to the plate. A bioassay on our chip is prepared by pipetting the (immuno)reactants on the inlet port and then incubating them in the microchannel at zero flow velocity. Therefore, the interaction between molecules from the liquid phase and the solid surface relies solely on diffusion. Nevertheless, the presence of bound antigen is measured by means of single-potential amperometry where a flow is needed to transport the H2 O2 generated by GOx-avidin to the working electrode (Fig. 4a and b). An injection flow-like method was used to amplify the amperometric signal, where glucose solution is pumped at a constant flow rate through the chip and stopped for a certain time interval, allowing the H2 O2 to accumulate in the sensor capillary. Subsequent restart of the pump yields a signal peak (Fig. 4c) caused by the H2 O2 oxidation at the working electrode at +350 mV. Higher signals and their improved separation were observed with increasing stop-flow intervals. Four min was established as an optimal read-out time for the cTnI assay, however, the incubation interval can be adjusted accordingly to the particular assay performance. The analytical performance of the cTnI assay is shown for both immobilization strategies, passivating (Fig. 5a and b) and direct (Fig. 5c), additionally compared with standard benchtop ELISA (Fig. 5d). Fig. 5a depicts the example of data acquisition and assessment, where the height of the anodic peaks (Imax ) was used to calculate the respective points of the calibration curve plotted. The peak area is representative of the cTnI concentration range as well, however, it fails to provide robust data in comparison to Imax (data not shown). Standard four-parameters logistic fit was used for curve-fitting analysis. As can be seen in Fig. 5b and c and Table 1, the type of immobilization method strongly influenced the assay characteristics. The direct immobilization method is identical with that used in our
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
483
Fig. 4. Schematic representation of the measurement setup and an example of the electrochemical data acquisition measured with the BPEI modified immunochip. (a) Current measurement setup enables four chips to be measured simultaneously. The flow of liquids through the chips is generated by the syringe pump in withdrawal mode. (b) Illustration of the enzymatic H2 O2 generation in the immobilization capillary and its subsequent transport and oxidation at the working electrode. (c) Example of the amperometric measurement of two cTnI concentrations showing the oxidation peaks recorded after one, two and four min stop-flow time intervals.
previous reports and it is a good reference point for a comparison study. The on-chip cTnI assay was less precise with coefficients of variations reaching the last acceptable 20% intra-assay value throughout the calibration, with a limit of detection (LOD) of 0.45 ng ml−1 , which is a 22-fold decrease in comparison to the
conventional cTnI assay. LOD is taken here as the signal obtained from a control (absent analyte) sample plus three times the standard deviation of the control. Also, the direct immobilization method showed an increased signal to the control sample (Fig. 5c). These findings may be attributed to the low packing density and
Fig. 5. cTnI calibration on BPEI coated chips represented by (a) oxidation peaks recorded with 4 min stop-flow and (b) plotted as peak height (Imax ) normalized with 4parameter logistic fit. Comparison with the same cTnI assay using (c) direct on-chip immobilization method and (d) conventional microtiter plate with optical detection. Error bars represent the SD and are collected from three parallel measurements intra-assay.
484
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
activity of immobilized 19C7/16A11 mAbs in combination with higher molecular weight of the analyte molecule. In the case where the analyte exceeds the spacing between immobilized antibodies, steric hindrance between antigen molecules may prevent all of the antibody active sites from being utilized. Using the BPEI passivating method, the cTnI assay in undiluted serum performed with an LOD of 0.025 ng ml−1 . Although the intra-assay precision for lower cTnI concentrations did not exceed 12% CV, for the highest measured cTnI concentration, it reached 20% CV. Taking into account all measurement parameters, the BPEI coating showed a significant improvement of conjugation efficiency and surface activity of the attached antibodies together with good protein-resistance properties. The amperometric response of both immobilization strategies were comparable, nevertheless the passivating method utilized 1 g ml−1 of the GOx-avidin label and 4 min stop-flow incubation, whereas the direct method required 4 g ml−1 of the GOx-avidin and 10 min stop-flow incubation. On the other hand, the higher packing density of BPEI-immobilized antibodies may explain the increased saturation behavior and the lowered inflexion point (IC50 ) observed with higher cTnI concentrations. Direct comparison of tested immobilization strategies in terms of conjugation efficiency is problematic due to their very different performance, as can be seen in Fig. 3. When the same concentration of GOx was used, we obtained approximately 40-times and 60-times higher response using LPEI and BPEI respectively, in comparison to the direct immobilization method (data not shown).
has more than doubled, a significant improvement, which should benefit other research areas, but remains insufficient for long-term storage required for POC diagnostics. Compared to the conventional microplate ELISA, the utilization of the microfluidic chip accelerates the assay preparation protocol by a factor of two and requires only 5 l of immunoreactants, which substantially lowers the total assay costs. The utilization of sensitive amperometric detection shortens the signal read-out time to minutes and enables a high level of integration with the platform. The flexibility in design, low-cost production and ease of handling, both the platform and presented immobilization strategies will enable many potential bioassay applications.
Acknowledgments The authors would like to thank the University Medical Center Freiburg for providing serum samples and the European Commission for financial support under the framework of Marie Curie Research and Trainings Network “Cellcheck” (MCRTN-CT-2006035854).
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.snb.2014.12.006.
References 4. Conclusion Many biosensing platforms are fabricated on rigid materials ignoring the relevance of their biofunctionalization and biomolecule fate. On the other hand, many immobilization techniques are not compatible with the target platform architecture or material composition. The motivation behind this study was to develop an efficient immobilization method for convenient biomolecule attachment in the context of immunosensors by inserting a “passivating” layer of PEI between the biomolecule and the surface. The primary criteria were the applicability of the surface modification to virtually any platform design with focus on microfluidics and the construction of scalable, functionally controllable biomolecular surface. We demonstrated the applicability of the disposable microfluidic chip to a wide range of immunoassay formats and diverse analytes in clinically relevant samples. Herein, the optimized cTnI assay utilized two capture mAbs 19C7/16A11 and one detection mAb MF4 to maximize the cardiac specificity and sensitivity while maintaining the highest signal possible. The LOD of 18.5 pg ml−1 and imprecision ≤12% CV was measured on the microtiter plate. The chip performance was assessed with two immobilization approaches. The direct method based on EDC/SNHS coupling yielded poor assay characteristics. Both PEI-based procedures were extremely powerful since they could immobilize large amounts of biomolecules while showing higher biological activity in comparison to the EDC/SNHS coupling. The adsorbed LPEI showed potential due to its simplicity and compatibility with various materials; however, failed to provide sufficient stability (approx. two days) against a high number of washing steps in comparison to immobilized BPEI (approx. six days). In contrast, the BPEI modified chip allowed reaching an LOD of 25 pg ml−1 with 10% CV, with other assay parameters comparable to the benchtop ELISA. The main merit of BPEI surface modification will strongly depend on its application; we have observed up to 60-times higher biomolecule loading/surface activity and approximately three-times lower nonspecific adsorption. Additionally, the stability of attached probes
[1] T. Porstmann, S.T. Kiessig, Enzyme immunoassay techniques an overview, J. Immunol. Methods 150 (1992) 5–21. [2] P. Tijssen, A. Adam, Enzyme-linked immunosorbent assays and developments in techniques using latex beads, Curr. Opin. Immunol. 3 (1991) 233–237. [3] C.D. Chin, T. Laksanasopin, Y.K. Cheung, D. Steinmiller, V. Linder, H. Parsa, et al., Microfluidics-based diagnostics of infectious diseases in the developing world, Nat. Med. 17 (2011) 1015–1019. [4] S.R. Narayanan, L.J. Crane, Affinity chromatography supports: a look at performance requirements, Trends Biotechnol. 8 (1990) 12–16. [5] P. Cutler, Affinity chromatography, in: Protein Purification Protocols, Springer, Totowa, New Jersey, 1996, pp. 157–168. [6] T.I. Yin, Y. Zhao, J. Horak, H. Bakirci, H.H. Liao, H.H. Tsai, et al., A microcantilever sensor chip based on contact angle analysis for a label-free troponin I immunoassay, Lab Chip 13 (2013) 834–842. [7] X. Fan, I.M. White, S.I. Shopova, H. Zhu, J.D. Suter, Y. Sun, Sensitive optical biosensors for unlabeled targets: a review, Anal. Chim. Acta 620 (2008) 8–26. [8] P. Esser, Activity of adsorbed antibodies, Appl. Note 11b (1997). [9] J.N. Herron, H.K. Wang, V. Janatova¡, J.D. Durtschi, D.A. Christensen, K.D. Caldwell, et al., Orientation and activity of immobilized antibodies, in: Biopolymers at Interfaces, second ed., 2003, pp. 115–164. [10] D. Kim, A.E. Herr, Protein immobilization techniques for microfluidic assays, Biomicrofluidics 7 (2013) 041501. [11] A. Subramanian, S.J. Kennel, P.I. Oden, K.B. Jacobson, J. Woodward, M.J. Doktycz, Comparison of techniques for enzyme immobilization on silicon supports, Enzyme Microb. Technol. 24 (1999) 26–34. [12] P. Jonkheijm, D. Weinrich, H. Schröder, C.M. Niemeyer, H. Waldmann, Chemical strategies for generating protein biochips, Angew. Chem. 47 (2008) 9618–9647. [13] W.D. Oliveira, W.G. Glasser, Hydrogels from polysaccharides. I. Cellulose beads for chromatographic support, J. Appl. Polym. Sci. 60 (1996) 63–73. [14] B. Johnsson, S. Lofas, G. Lindquist, A. Edstom, R.-M.M. Hillgren, A. Hansson, Comparison of methods for immobilization to carboxymethyl dextran sensor surfaces by analysis of the specific activity of monoclonal antibodies, J. Mol. Recognit. 8 (1995) 125–131. [15] R. Wong, H. Tse, Lateral Flow Immunoassay, Springer, 2009. [16] J. Horak, C. Dincer, H. Bakirci, G. Urban, A disposable dry film photoresist-based microcapillary immunosensor chip for rapid detection of Epstein-Barr virus infection, Sens. Actuators B: Chem. 191 (2014) 813–820. [17] P. Vicennati, A. Giuliano, G. Ortaggi, A. Masotti, Polyethylenimine in medicinal chemistry, Curr. Med. Chem. 15 (2008) 2826–2839. [18] R.R. Burgess, Use of polyethyleneimine in purification of DNA-binding proteins, Methods Enzymol. 208 (1991) 3–10. [19] M. Teramoto, H. Nishibue, K. Okuhara, H. Ogawa, H. Kozono, H. Matsuyama, et al., Effect of addition of polyethyleneimine on thermal stability and activity of glucose dehydrogenase, Appl. Microbiol. Biotechnol. 38 (1992) 203–208. [20] J. Yakovleva, R. Davidsson, M. Bengtsson, T. Laurell, J. Emneus, Microfluidic enzyme immunosensors with immobilised protein A and G using chemiluminescence detection, Biosens. Bioelectron. 19 (2003) 21–34.
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485 [21] Y. Bai, C.G. Koh, M. Boreman, Y.J. Juang, I.C. Tang, L.J. Lee, et al., Surface modification for enhancing antibody binding on polymer-based microfluidic device for enzyme-linked immunosorbent assay, Langmuir 22 (2006) 9458–9467. [22] M.M. Andersson, R. Hatti-Kaul, Protein stabilising effect of polyethyleneimine, J. Biotechnol. 72 (1999) 21–31. [23] M. Erol, H. Du, S. Sukhishvili, Control of specific attachment of proteins by adsorption of polymer layers, Langmuir 22 (2006) 11329–11336. [24] J. Horak, C. Dincer, H. Bakirci, G. Urban, Sensitive, rapid and quantitative detection of substance P in serum samples using an integrated microfluidic immunochip, Biosens. Bioelectron. 58 (2014) 186–192. [25] B. McDonnell, S. Hearty, P. Leonard, R. O’Kennedy, Cardiac biomarkers and the case for point-of-care testing, Clin. Biochem. 42 (2009) 549–561. [26] A.H.B. Wu, Cardiac Markers, Humana Press Inc., 2003. [27] C.S. Farah, F.C. Reinach, The troponin complex and regulation of muscle contraction, FASEB J. 9 (1995) 755–767. [28] S.J. Maynard, I.B.A. Menown, A.A.J. Adgey, Troponin T or troponin I as cardiac markers in ischaemic heart disease, Heart 83 (2000) 371–373. [29] J. Sarko, C.V. Pollack Jr., Cardiac troponins, J. Emerg. Med. 23 (2002) 57. [30] B. Cummins, M.L. Auckland, P. Cummins, Cardiac-specific troponin I radioimmunoassay in the diagnosis of acute myocardial infarction, Am. Heart J. 113 (1987) 1333–1344. [31] A. Katrukha, A. Bereznikova, V. Filatov, T. Esakova, Biochemical factors influencing measurement of cardiac troponin I in serum, Clin. Chem. Lab. Med. 37 (1999) 1091–1095. [32] T. Keller, T. Zeller, F. Ojeda, S. Tzikas, L. Lillpopp, C. Sinning, et al., Serial changes in highly sensitive troponin I assay and early diagnosis of myocardial infarction, J. Am. Med. Assoc. 306 (2011) 2684–2693. [33] J.P. Chapelle, M.C. Aldenhoff, L. Pierard, J. Gielen, Comparison of cardiac troponin I measurements on whole blood and plasma on the stratus CS analyzer and comparison with AxSYM, Clin. Chem. 46 (2000) 1864–1866. [34] E. Giannitsis, H.A. Katus, 99th percentile and analytical imprecision of troponin and creatine kinase-MB mass assays: an objective platform for comparison of assay performance, Clin. Chem. 49 (2003) 1248. [35] M. Panteghini, F. Pagani, K.T.J. Yeo, F.S. Apple, R.H. Christenson, F. Dati, et al., Evaluation of imprecision for cardiac troponin assays at low-range concentrations, Clin. Chem. 50 (2004) 327. [36] J.R. Tate, Troponin revisited 2008: assay performance, Clin. Chem. Lab. Med. 46 (2008) 1489–1500. [37] J. Wang, Electrochemical glucose biosensors, Chem. Rev. 108 (2008) 814–825. [38] G.T. Hermanson, Bioconjugate Techniques, second ed., Elsevier Science, 2008.
485
[39] I. Migneault, C. Dartiguenave, M.J. Bertrand, K.C. Waldron, Glutaraldehyde: behavior in aqueous solution, reaction with proteins, and application to enzyme crosslinking, BioTechniques 37 (2004) 790–806.
Biographies Josef Horak received his master’s degree in Biotechnology from the Brno University of Technology. This was followed by a one year scholarship at the Technical University of Lisbon, where he worked on cell based sensors. In 2013, he received his Ph.D. from the University of Freiburg, Germany, Department of Microsystems Engineering, where he is currently working as a group leader in the field of bioanalytical microsystems with a focus on immunosensors. Can Dincer received his master’s degree in Microsystems Engineering from the University of Freiburg in 2009, where he later worked as a scientific staff member at the Laboratory for Sensors. Since 2013, Can Dincer is head of the Nanosensors group in the field of immunosensors, microtechnology and sensor applications in medicine. Edvina Qelibari works as a student assistant at the Department of Microsystems Engineering at the University of Freiburg. Huseyin Bakirci studied Biochemistry at the University Louis Pasteur in France. In 2006, he obtained his Ph.D. in chemistry from the Department of Physical Chemistry, University of Basel. He continued at the same University as a postdoctoral researcher in the Department of Chemistry and Biochemistry and later on at the Division of Nephrology and Hypertension at the University Hospital of Bern in Switzerland. From 2007 till 2013, he was a group leader of bioanalytical microsystems at the Department of Microsystems Engineering, University of Freiburg. Gerald Urban studied technical physics at the TU Vienna, where he subsequently obtained his Ph.D. from the Department of Electrical Engineering. In 1985, he cofounded the company OSC in Cleveland, Ohio and in Vienna. He finished his post-doctoral studies in 1987 at the neurophysiological department in Münster, Germany. From 1990 till 2002 he was scientific director of the Ludwig Boltzmann Institute of Biomedical Microengineering in Vienna. He finished his habilitation in 1994 and became a full professor at the University of Freiburg in 1997. From 1999 to 2002 he was dean of the Faculty for Applied Science at the University of Freiburg.
Contents lists available at ScienceDirect
Sensors and Actuators B: Chemical journal homepage: www.elsevier.com/locate/snb
Polymer-modified microfluidic immunochip for enhanced electrochemical detection of troponin I Josef Horak a,∗ , Can Dincer a,b , Edvina Qelibari a , Hüseyin Bakirci a , Gerald Urban a,b a b
Laboratory for Sensors, Department of Microsystems Engineering (IMTEK), University of Freiburg, Georges-Köhler-Allee 103, 79110 Freiburg, Germany Freiburg Materials Research Center (FMF), University of Freiburg, Stefan-Meier-Straße 21, 79104 Freiburg, Germany
a r t i c l e
i n f o
Article history: Received 16 October 2014 Received in revised form 29 November 2014 Accepted 2 December 2014 Available online 11 December 2014 Keywords: Immunochip Immunoassay Electrochemical detection Polyethylenimine Troponin I
a b s t r a c t We report enhanced electrochemical detection of a cardiac biomarker using a combination of a disposable microfluidic immunochip fabricated in Vacrel® 8100 photoresist film and a highly effective surface functionalization employing polyethylenimine (PEI). The use of the photoresist film enables fast prototyping and low-cost production without the need for a cleanroom. We use the surface carboxylates of the photoresist to biofunctionalize the microchannel on the chip using direct amine-specific coupling and modification via adsorbed and immobilized PEI in both linear (LPEI) and branched (BPEI) form. Characteristics of each immobilization strategy are assessed by a sandwich immunoassay for troponin I quantification in serum. The best assay performance is achieved using the immunochip modified with immobilized BPEI: the antigen was routinely detected at concentrations of 25 pg ml−1 in 4 min read-out time and 5 l serum sample, representing an 18-fold improvement of the detection limit and 2.5-times faster read-out time in comparison to the assay implemented over amine-reactive esters without the PEI coating. We demonstrate that immobilized BPEI represents a stable and tunable scaffold that enhances biomolecule immobilization up to 60-times and stability 2-fold while lowering non-specific binding by factor of three. This system provides a versatile and nonaggressive means for incorporation of biological material to virtually any platform design. © 2014 Elsevier B.V. All rights reserved.
1. Introduction The development of a versatile immunosensing platform that can effectively and economically detect a broad range of different molecular targets is highly desirable. Performance characteristics of the vast majority of immunosensors, and biosensors in general, are directly linked to the choice of materials and applicable immobilization methods to attach the immunoreactants onto solid surfaces. The ideal immobilization strategy should employ mild chemical conditions and allow large quantities of biomolecules to be immobilized while retaining their biological activity. The immobilization surface should provide a large area for antibody–antigen interaction within a small total volume and should limit the non-specific protein adsorption. In terms of practical use, immobilization methods should extend the shelf life, which is a necessary criterion for commercial biosensors such as point-of-care (POC) diagnostics. Without question, polystyrene, silica and polysaccharides are the three most common substrates for biomolecule
∗ Corresponding author. Tel.: +49 0761 203 7264; fax: +49 0761 203 7262. E-mail address: [email protected] (J. Horak). http://dx.doi.org/10.1016/j.snb.2014.12.006 0925-4005/© 2014 Elsevier B.V. All rights reserved.
attachment. Polystyrene is the traditional substrate for enzymelinked immunosorbent assays (ELISA) in the form of microtiter plates or beads [1,2], but has been employed in biosensors as well [3]. Silica can be fabricated into a number of different forms including particles, wafers and thin films deposited by plasma chemical vapor deposition. Silica particles are employed in both chromatography and immunoassays [4,5], while silica wafers and thin films are increasingly being used in biosensors [6,7]. The same issues arise in both traditional immunoassays and biosensors: the decreased surface activity of the immobilized antibodies (or biomolecules in general) and non-specific binding (NSB) of various sample components to the solid phase. Antibodies attached by passive adsorption to polystyrene undergo conformational changes and remain only marginally active (<10%) [8,9]. The covalent immobilization used for silica and for PDMS relies predominantly on silanization protocols [10], which is a time consuming procedure and requires aggressive conditions that are not always compatible with non-silica parts of the packaging or housing [11]. These can have a negative impact on both the assay sensitivity and NSB – the former because relatively few active capture molecules are available to bind the analyte, and the latter because sample components are inclined to interact nonspecifically with the
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
preponderance of inactive antibodies, which may be partially denatured. To tackle these problems, various polymers have been used for biosensor applications. Instead of attaching proteins onto a twodimensional solid surface, biomolecules can diffuse into a porous matrix formed by polymer membranes or surface coatings [12]. Polysaccharides were employed in a number of different types of chromatographic supports and immobilization matrices in biosensors such as carboxymethyl dextran in Biacore systems [13,14]. The traditional membrane material for the vast majority of lateral flow assays is nitrocellulose; nylon and polyvinylidene fluoride membranes had only limited success due to various factors such as high cost and limited utility [12,15]. Based on the aforementioned considerations, we merged the above approaches into a single platform to simultaneously take advantage of our microfluidic immunochip [16] and a novel immobilization strategy. We combined an microfabrication strategy using the flexible dry film photoresist, Vacrel 8100® , which offers good wetting properties for passive fluid control and easy biofunctionalization, together with an immobilization strategy in which the surface is first coated with a thin film of polyethylenimine that acts as a matrix for immobilization of biomolecules. We will refer to such a polymer coating as a “passivating” layer since it limits the interaction of proteins and other sample components with the surface – a fitting term used by Herron at al. [9]. Polyethylenimines are polycationic aliphatic polymers that can be produced either in linear form (LPEI) or with varying degrees of branching (BPEI). PEI has a high density of amino groups – every third atom in the polymer chain is a nitrogen atom. BPEI contains primary, secondary and tertiary amino groups with a ratio of 1:2:1, while LPEI mostly contains only secondary amines. In general, PEIs are both highly basic and water soluble polymers available in molecular weights ranging from 700 Da to 1000 kDa. They have been extensively studied predominantly in medicinal chemistry [17], but also for non-pharmaceutical use [18,19]. Importantly for our application, the polymer has been shown to be effective for immobilization of biocatalysts to solid supports by adsorption [20,21] and to have a positive effect on enzyme activity and stability [22] while having unique protein-resistant properties [23]. In our previous report, we demonstrated the high level of performance of our immunosensing platform using a competitive ELISA for substance P (an analyte of low molecular weight) and simple carbodiimide-based coupling chemistry [24]. To assess the benefit of the PEI modified immunosensor, we present here rapid detection with clinically relevant sensitivity of troponin I (cTnI) in a sandwich assay format. cTnI is the “gold standard” for diagnosis of cardiac muscle cell damage and death [25]. cTnI (23.8 kDa) is a component of the troponin ITC complex (77.2 kDa), further comprising troponin T (35 kDa) and troponin C (18.4 kDa), a heteromeric protein that plays an important role in the regulation of skeletal and cardiac muscle contraction [26,27]. cTnI is expressed only in myocardium and there are no examples known of cTnI expression in healthy or injured skeletal muscle or in other tissue types [28,29], hence it is currently widely used for the diagnosis of acute myocardial infarction, unstable angina, post-surgery myocardium trauma, as well as several other diseases related to cardiac muscle injury [30].
2. Material and methods
479
from HyTest Ltd., Finland. Serum and blood samples were obtained from University Medical Center Freiburg. Glucose oxidase avidin conjugate (GOx-avidin) was obtained from Biomol, Germany. All other chemicals were purchased from Sigma–Aldrich or otherwise as stated in the text. 2.2. The chip fabrication A detailed description of the wafer fabrication process utilizing Vacrel® 8100 can be found in our previous report [16]. 2.3. Microtiter plate-based troponin I ELISA See S1, Supplementary material for further information. Every step of the immunoreactions was followed by three washing cycles on a commercial ELISA microplate washer using wash buffer (10 mM PBS, 138 mM NaCl, 2.7 mM KCl, 0.005% Tween 20, pH 7.4). Standard coating volume of 100 l per well was used for all immunoreactants, 200 l was used for blocking. Plates (NuncImmunoTM LockWellTM Modules, Nunc GmbH & Co. KG, Germany) were incubated sealed on a plate shaker preheated to 25 ◦ C and at 450 rpm. Immunoreactants were diluted in PBS buffer (10 mM PBS, 138 mM NaCl, 2.7 mM KCl, pH 7.4). Microtiter plate was coated with capture antibody combination, 16A11 and 19C7, to a final concentration of 10 g ml−1 diluted in PBS. After 2 h incubation, the plate was washed. To saturate any binding sites not already occupied by the capture antibody, incubation with 200 l of blocking solution was carried out for 30 min at 25 ◦ C on a microplate shaker followed by overnight incubation at 4 ◦ C without shaking. After subsequent washing, the plate was ready to use. Troponin ITC complex was spiked into cTnI-free serum to concentration range 0.02–12.5 ng ml−1 using a 5-fold serial dilution. The calibrators were incubated for 2 h followed by a washing step and 1 h incubation with 4 g ml−1 of biotinylated detection antibody, MF4b. 100 l of a solution of GOx conjugated avidin diluted to 1 g ml−1 in reagent diluent was added and the plates were incubated for 30 min at 25 ◦ C. After subsequent washing, 50 l of ABTS diluted in phosphate-citrate buffer (200 mM Na2 HPO4 , 100 mM citric acid, pH 5.6) to concentration 2.2 mg ml−1 were pipetted into each well. Solutions containing 160 mM of glucose and 0.8 mg ml−1 of horseradish peroxidase both in phosphatecitrate buffer were mixed in 1:1 ratio. A volume of 50 l of this solution was pipetted into each well. The plate was incubated without shaking for 1 h at 25 ◦ C and then immediately measured on an absorbance microplate reader (SpectraMax 340PC384, Molecular Devices GmbH, Germany) preheated at 25 ◦ C at 405 nm. Software SoftMax Pro (Molecular Devices GmbH, Germany) was used to collect the data. 2.4. Chip-based troponin I ELISA For every incubation step, a 5 l of (immuno)reagents was pipetted on the chip inlet to fill the chip immobilization capillary. Chips were incubated in closed Petri dishes at 25 ◦ C. Every step of the immunoreactions was followed by a washing step with a custom made vacuum pump using 200 l of a wash buffer. PBS buffer and wash buffer were identical to those used for microtiter plate ELISA. Methods for all immobilization strategies tested are described in S2, Supplementary material.
2.1. Materials and chemicals Pyralux® AP, Vacrel® 8100 and Teflon® 1600 AF were purchased from DuPontTM . Monoclonal capture and detection antibodies (mAb); troponin ITC complex and cTnI-free serum were purchased
2.4.1. Direct immobilization strategy The immobilization area of the chip was pretreated for 10 min with 10% Na2 CO3 and washed with 100 mM 2-(Nmorpholino)ethanesulfonic acid (MES) buffer, pH 6.0. The
480
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
capillary was treated with activation buffer (100 mM 1-ethyl-3-[3dimethylaminopropyl]carbodiimide hydrochloride (EDC), 200 mM N-hydroxysulfosuccinimide (SNHS) solution buffered at pH 6.0 with 100 mM MES, 0.9% NaCl) for 60 min. A 50 g ml−1 solution of 19C7 and 16A11 mAbs were prepared in PBS buffer and incubated for 60 min in the activated capillary. After a washing step, the capillary was blocked for 30 min using 1% BSA in PBS buffer. The rest of the assay workflow was identical with the conventional cTnI ELISA described above, except for the higher utilized concentration of the GOx-avidin label (4 g ml−1 ) and duration of incubation steps cut by half. 2.4.2. Passivating layer method utilizing immobilized BPEI The immobilization area of the chip was pretreated for 10 min with 10% Na2 CO3 and washed with 100 mM MES buffer, pH 6.0. The capillary was treated with activation buffer (100 mM EDC, 200 mM sulfo-NHS solution buffered at pH 6.0 with 100 mM MES, 0.9% NaCl) for 60 min. 5% (w/v) solution of BPEI was prepared in PBS buffer, vortexed well and neutralized with 6 M HCl to pH 8. After 60 min incubation at 25 ◦ C in the activated capillary, the channel was washed with 200 l of 1 M NaCl in ultra-pure H2 O and 200 l of PBS. The BPEI was reacted with 2.5% glutaraldehyde (GA) solution containing 10 mM NaCNBH3 for 1 h and washed with 200 l of ultra-pure H2 O and 100 l of 50 mM carbonate–bicarbonate buffer, pH 9.6. A 5 g ml−1 solution of 19C7 and 16A11 mAbs were prepared in 50 mM carbonate–bicarbonate buffer, pH 9.6 and incubated in the activated channel for 60 min. Chips were washed with 200 l of 50 mM carbonate–bicarbonate pH 9.6 and incubated for 30 min with 25 mM NaCNBH3 prepared in PBS buffer. After washing with 100 l of PBS, chips were blocked with 1% BSA blocking solution containing 25 mM of NaCNBH3 for 30 min followed by washing. The rest of the assay workflow was identical with the conventional cTnI ELISA described above, except for the duration of incubation steps cut by half. 2.4.3. Amperometric measurement Chips were fastened into a custom-made adapter connected to a potentiostat (Jobst Technologies, Germany) and a syringe pump (PHD2000 Harvard Aparatus; syringes Hamilton Company); constant flow (withdraw mode) of 15 l min−1 was used throughout the measurements. 100 mM PBS, 100 mM NaCl, pH 7.4 buffer was used to establish the baseline, glucose solution substrate (40 mM glucose in 100 mM PBS, 100 mM NaCl, pH 7.4) was used to generate the electrochemical signal. The measurement was recorded with bioMON software (Jobst Technologies, Germany) at +350 mV vs. Ag/AgCl on-chip pseudo-reference electrode. 3. Results and discussion In our previous reports, we have shown the high degree of performance of our immunosensing platform using qualitative direct ELISA for Epstein-Bar virus detection [16] and quantitative competitive ELISA for small peptide substance P [23]. To complete the demonstration that the immunochip is a versatile platform applicable to all of the major analyte classes, we investigated whether it could detect protein biomarkers, using troponin I as a model system. 3.1. Immunochip design and fabrication As can be seen in Fig. 1a, a single chip consists of a microchannel, an electrochemical measurement cell containing Pt working, counter and Ag/AgCl pseudo-reference electrodes (Fig. 1b), and
contact pads. The immobilization area of the channel (Fig. 1c), bordered by the inlet port and the stopping barrier, possesses a surface of 112 mm2 and 3.6 l volume, which creates a high surface to volume ratio of 310 cm−1 . Detailed geometry and dimensions of the electrodes and microfluidic channel are depicted in Fig. S3. For the bioassay preparation, all reagents are pipetted onto the inlet and enter the immobilization channel by capillary force, so the reagents coat a well-defined area and do not contaminate the electrochemical measurement cell. The flexible microfluidic chip is produced completely on waferlevel by a standard photolithographic technique employing dry film resist Vacrel® 8100. Briefly, this fabrication strategy relies on lamination of the film photoresist to a platinum patterned polyimide wafer using heat and pressure. All chip structures are subsequently generated by photolithography. A single 6-inch wafer comprises 40 chips. The most apparent advantages of using dry film resist are fast processing and prototyping requiring minimal material and device costs; except for Pt deposition, all fabrication steps were performed outside of a cleanroom. The less apparent advantage of Vacrel® 8100 is the possibility of using its surface carboxylates for amine-specific “direct” biofunctionalization. 3.2. Development of troponin I ELISA The crucial issue of the cTnI assay performance is the choice of an antibody combination that would recognize the cTnI epitopes that are the least affected by (i) conformational changes and (ii) the complexation of the antigen when released into the patient’s blood from necrotic tissue after acute myocardial infarction [31]. The developed cTnI ELISA is based on the standard sandwich assay format. Two mouse monoclonal antibodies (mAb), 19C7 and 16A11, binding epitopes located in the central stable part of the cTnI molecule (between 30 and 110 amino acid residues) were used for solid phase immobilization. A third high affinity mAb conjugated biotin, MF4b (recognizing epitope 190–196), was used for detection. In the last step, the sandwich complex was labeled with a GOx conjugated avidin. The subsequent implementation of the cTnI ELISA on the chip and using electrochemical detection were ample reasons to develop the assay with GOx instead of the more common horseradish peroxidase label. The assay concept and antibody combination were first optimized on a microtiter plate before being implemented on the chip (S1, Supplementary material). Since the cTnI assays rely on serum troponin level [32,33], undiluted cTnI-free serum was used to evaluate the assay parameter both on the plate and the immunochip. However, measurements performed with spiked whole blood and serum samples showed statistically indistinguishable values of cTnI in the entire calibration range (Fig. S2). International guidelines endorsed total cTnI assay imprecision at their 99th percentile reference value to be ≤10% coefficient of variation (CV) [34]. In our setting, the assessment of the 99th percentile was not possible (ranges between 0.01 and 0.3 ng ml−1 for commercial assays) [35,36]. Our assay performed with 11.6% CV at 0.02 ng ml−1 in serum; diagnostic levels of cTnI are considered negative up to 0.1 ng ml−1 . 3.3. On-chip implementation of troponin I ELISA We experimented with five different strategies to immobilize the biological probes on Vacrel® 8100 which composes the immobilization area of the immunochip capillary. All methods employed random covalent coupling via primary amines present predominantly on lysine residues of the biomolecules; these included: (i) direct covalent coupling to the photoresist material via carboxyl-to-amine crosslinking, (ii) covalent attachment to adsorbed LPEI, (iii) covalent attachment to immobilized LPEI, (iv) covalent attachment to adsorbed BPEI and
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
481
Fig. 1. Summary of the cTnI ELISA implementation on the BPEI modified immunochip. (a) Photograph of the utilized immunochip (image without front cover) showing the main dimensions and functional elements. (b) The electrochemical measurement cell containing platinum working (WE), counter (CE) and Ag/AgCl reference (RE) electrodes. (c) The microfluidic channel. The surface modification and assay preparation takes place in the immobilization capillary (highlighted gray) bordered by the inlet and stopping barrier. (d) Direct immobilization strategy and (e) passivating layer method utilizing immobilized BPEI are shown.
Fig. 2. Comparison of immobilization approaches tested by performing on-chip cTnI assay. Signals represent 4 min stop-flow. Signal-to-background ratio (SBR) was calculated for measurements where positive signals were higher than a control sample. Error bars represent SD and are collected from three parallel measurements.
finally (v) covalent attachment to immobilized BPEI. Alongside the cTnI assay, the immobilization methods were evaluated using GOx as a model biomolecule. This enzyme is well characterized and commonly used for biosensing applications including ours [37]. Further text will be restricted only to the comparison of the direct immobilization strategy (Fig. 1d) with the passivating layer method utilizing covalently attached BPEI (Fig. 1e). All other approaches (S2, Supplementary material) failed to provide satisfactory results (Fig. 2) except for covalent protein attachment to adsorbed LPEI. The reason for this was the generally lower stability of the adsorbed PEIs to the high number of washing steps required for immunoassay preparation and in the long-term, since the PEI-Vacrel® interactions are predominantly electrostatic in nature. As shown in S4, Supplementary material, the immobilized BPEI was stable for six days while adsorbed LPEI only for two days. Nevertheless, the immobilization strategy via adsorbed LPEI showed potential for the efficient attachment of biomolecules in applications that do not involve such a high number of washing steps.
3.3.1. Direct immobilization method Briefly, the utilized photoresist formulation contains acrylic binders and various multifunctional acrylates possessing carboxylic and carboxylate groups. These can be reacted with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide and Nhydroxysulfosuccinimide to form an reactive O-acylisourea and succinimidyl ester intermediates that are easily displaceable by nucleophilic attack from primary amines of the target protein. SNHS ester is considered to be more stable than the O-acylisourea intermediate while allowing for efficient conjugation to primary amines at physiologic pH. The direct immobilization strategy represents a fast and convenient way of biomolecule attachment, nevertheless, requires a higher concentration of the probe. The optimized on-chip cTnI assay utilized higher coating concentration of the capture mAbs combination (50 g ml−1 ) and the GOx-avidin label (4 g ml−1 ) in comparison to the microplate assay. BSA was used to simultaneously block the unreacted crosslinker intermediates and as a subsequent blocking step. Storage of such pre-immobilized and pre-blocked chips strongly depended on the stability of the immobilized biomolecules. Stored at 4 ◦ C, blocked chips containing 19C7/16A11 mAbs had to be used within 24 h; in the case of GOx, chips had to be used immediately upon preparation.
3.3.2. Passivating layer method using immobilized BPEI In this immobilization strategy, the cTnI specific capture mAbs were coupled to the capillary surface coated with BPEI (25,000 Mw). Our choice was motivated by several factors: (i) PEIs of higher molecular weight are relatively inert to protein adsorption [23] and for some proteins, e.g. glucose oxidase, can resist the non-specific adsorption even without a dedicated blocking strategy (S4, Supplementary material). (ii) PEIs can be readily chemically functionalized with proteins of interest. (iii) Both the protein blocking properties and functionalization efficiency can be optimized by the PEI concentration utilized. The primary and secondary amino groups of BPEI were used both for immobilizing the polymer to the EDC/SNHS activated channel surface and for conjugating the biomolecules to the BPEI via glutaraldehyde.
482
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485 Table 1 cTnI assay parameters for chip-based direct and passivating immobilization strategies compared to the conventional microtiter plate. Parameters
Chip (direct)
Chip (passivating)
Microtiter plate
LOD (ng ml−1 ) LOQ (ng ml−1 ) Max. imprecision (%CV) Imprecision at LOD (%CV) IC50 (ng ml−1 ) Slope (nA ng ml−1 ) Total assay time (h) Reagent/sample volume (l)a
0.45 0.75 ≤20 ≤14 85.4 0.62 5
0.025 0.042 ≤20 ≤12 6.36 1.01 6
0.019 0.03 ≤12 ≤12 20.62 1.21 6 + overnightb 100/200c
a b c
Fig. 3. Comparison of immobilization/surface activity and stability of biomolecules (GOx) coupled on the immunochip via direct and BPEI immobilization strategies. Signals represent 5 min stop-flow. Error bars represent SD and are collected from three parallel measurements.
Prior to the BPEI immobilization, the basic BPEI solution (pH 10.5) had to be neutralized to pH 8 in order to prolong the stability of the succinimidyl ester, which is the amine-reactive intermediate of the activated surface carboxylates. The half-life of the succinimidyl ester at this pH value is approximately 1 h, providing enough time for successful BPEI immobilization. However, already at pH 8.6, the stability of SNHS ester decreases to only 10 min [38]. After the immobilization step, wash solution containing 1 M sodium chloride was used to remove the unbound BPEI from the immobilization capillary surface. The biofunctionalization of BPEI had to be optimized for different protein targets. The immobilization of antibodies on GA activated BPEI required utilization of immobilization buffer at pH 9.6 to deprotonate the lysine -amino group (pKa 10.5 in polypeptides). Moreover, we used an additional blocking step, simultaneously deactivating the unreacted GA aldehyde groups with BSA. However, for GOx conjugation, only a negligible difference in immobilization efficiency was observed for different buffer pH (pH 7.4 vs. 9.6). Although formation of an unstable Schiff base on both ends of the monomeric GA has been ruled out as a mechanism for GA crosslinking with proteins [39], it was necessary to use sodium cyanoborhydride to convert the Schiff base to a stable secondary amine. Use of the BPEI passivating layer enabled lowering the concentration of the immobilized capture antibodies (5 g ml−1 ) and prolonging the stability of the immobilized biomolecules. In comparison to the direct immobilization method, the use of a passivating layer extended the storage of pre-immobilized and preblocked chips to 48 h at 4 ◦ C. In the case of GOx, as can be seen in Fig. 3, the effect of PEI on long-term stability was significant; nevertheless, immobilized GOx lost about 64% of its activity (93% with direct immobilization) in 24 h when stored at 4 ◦ C. As mentioned earlier, covalent attachment of the PEI to the capillary surface prolongs its stability and usability to six days at 4 ◦ C. The above mentioned effect on prolonged biomolecule stability was significant and may be beneficial for a variety of applications performed in laboratory conditions where high biomolecule loading is of interest. Nevertheless, this immobilization strategy is not sufficiently stable for long-term storage, as may be required for commercial POC devices, such as lateral flow immunoassay systems using predominantly nitrocellulose as a protein immobilization matrix. Specifically, various qualitative tests for troponin T and troponin I in the form of sticks or cassettes are marketed with a
5
Volume per well/chip. Blocking step. Volume used for plate blocking.
declared stability of several weeks at ambient temperature. On the other hand, these traditionally designed lateral flow tests suffer from performance limitations; most notably sensitivity and reproducibility limits their applications in quantitative systems [15]. 3.4. The electrochemical immunoassay of troponin I To determine the effect of the immobilization strategies, we have established the assay parameters by performing cTnI calibration curves in undiluted human serum. Mimicking the physiological conditions, cTnI, in the form of human cardiac troponin ITC complex, was spiked into undiluted human serum over the concentration range 20 pg ml−1 to 12.5 ng ml−1 using a 5-fold serial dilution and pipetted on the preimmobilized chips as described in Section 2. After a washing step, MF4-biotin detection antibody and GOx-avidin were subsequently added, followed by a final washing. For all immunoreactions, we used 5 l of reagent volume and incubation times were cut to half in comparison to the plate. A bioassay on our chip is prepared by pipetting the (immuno)reactants on the inlet port and then incubating them in the microchannel at zero flow velocity. Therefore, the interaction between molecules from the liquid phase and the solid surface relies solely on diffusion. Nevertheless, the presence of bound antigen is measured by means of single-potential amperometry where a flow is needed to transport the H2 O2 generated by GOx-avidin to the working electrode (Fig. 4a and b). An injection flow-like method was used to amplify the amperometric signal, where glucose solution is pumped at a constant flow rate through the chip and stopped for a certain time interval, allowing the H2 O2 to accumulate in the sensor capillary. Subsequent restart of the pump yields a signal peak (Fig. 4c) caused by the H2 O2 oxidation at the working electrode at +350 mV. Higher signals and their improved separation were observed with increasing stop-flow intervals. Four min was established as an optimal read-out time for the cTnI assay, however, the incubation interval can be adjusted accordingly to the particular assay performance. The analytical performance of the cTnI assay is shown for both immobilization strategies, passivating (Fig. 5a and b) and direct (Fig. 5c), additionally compared with standard benchtop ELISA (Fig. 5d). Fig. 5a depicts the example of data acquisition and assessment, where the height of the anodic peaks (Imax ) was used to calculate the respective points of the calibration curve plotted. The peak area is representative of the cTnI concentration range as well, however, it fails to provide robust data in comparison to Imax (data not shown). Standard four-parameters logistic fit was used for curve-fitting analysis. As can be seen in Fig. 5b and c and Table 1, the type of immobilization method strongly influenced the assay characteristics. The direct immobilization method is identical with that used in our
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
483
Fig. 4. Schematic representation of the measurement setup and an example of the electrochemical data acquisition measured with the BPEI modified immunochip. (a) Current measurement setup enables four chips to be measured simultaneously. The flow of liquids through the chips is generated by the syringe pump in withdrawal mode. (b) Illustration of the enzymatic H2 O2 generation in the immobilization capillary and its subsequent transport and oxidation at the working electrode. (c) Example of the amperometric measurement of two cTnI concentrations showing the oxidation peaks recorded after one, two and four min stop-flow time intervals.
previous reports and it is a good reference point for a comparison study. The on-chip cTnI assay was less precise with coefficients of variations reaching the last acceptable 20% intra-assay value throughout the calibration, with a limit of detection (LOD) of 0.45 ng ml−1 , which is a 22-fold decrease in comparison to the
conventional cTnI assay. LOD is taken here as the signal obtained from a control (absent analyte) sample plus three times the standard deviation of the control. Also, the direct immobilization method showed an increased signal to the control sample (Fig. 5c). These findings may be attributed to the low packing density and
Fig. 5. cTnI calibration on BPEI coated chips represented by (a) oxidation peaks recorded with 4 min stop-flow and (b) plotted as peak height (Imax ) normalized with 4parameter logistic fit. Comparison with the same cTnI assay using (c) direct on-chip immobilization method and (d) conventional microtiter plate with optical detection. Error bars represent the SD and are collected from three parallel measurements intra-assay.
484
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485
activity of immobilized 19C7/16A11 mAbs in combination with higher molecular weight of the analyte molecule. In the case where the analyte exceeds the spacing between immobilized antibodies, steric hindrance between antigen molecules may prevent all of the antibody active sites from being utilized. Using the BPEI passivating method, the cTnI assay in undiluted serum performed with an LOD of 0.025 ng ml−1 . Although the intra-assay precision for lower cTnI concentrations did not exceed 12% CV, for the highest measured cTnI concentration, it reached 20% CV. Taking into account all measurement parameters, the BPEI coating showed a significant improvement of conjugation efficiency and surface activity of the attached antibodies together with good protein-resistance properties. The amperometric response of both immobilization strategies were comparable, nevertheless the passivating method utilized 1 g ml−1 of the GOx-avidin label and 4 min stop-flow incubation, whereas the direct method required 4 g ml−1 of the GOx-avidin and 10 min stop-flow incubation. On the other hand, the higher packing density of BPEI-immobilized antibodies may explain the increased saturation behavior and the lowered inflexion point (IC50 ) observed with higher cTnI concentrations. Direct comparison of tested immobilization strategies in terms of conjugation efficiency is problematic due to their very different performance, as can be seen in Fig. 3. When the same concentration of GOx was used, we obtained approximately 40-times and 60-times higher response using LPEI and BPEI respectively, in comparison to the direct immobilization method (data not shown).
has more than doubled, a significant improvement, which should benefit other research areas, but remains insufficient for long-term storage required for POC diagnostics. Compared to the conventional microplate ELISA, the utilization of the microfluidic chip accelerates the assay preparation protocol by a factor of two and requires only 5 l of immunoreactants, which substantially lowers the total assay costs. The utilization of sensitive amperometric detection shortens the signal read-out time to minutes and enables a high level of integration with the platform. The flexibility in design, low-cost production and ease of handling, both the platform and presented immobilization strategies will enable many potential bioassay applications.
Acknowledgments The authors would like to thank the University Medical Center Freiburg for providing serum samples and the European Commission for financial support under the framework of Marie Curie Research and Trainings Network “Cellcheck” (MCRTN-CT-2006035854).
Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.snb.2014.12.006.
References 4. Conclusion Many biosensing platforms are fabricated on rigid materials ignoring the relevance of their biofunctionalization and biomolecule fate. On the other hand, many immobilization techniques are not compatible with the target platform architecture or material composition. The motivation behind this study was to develop an efficient immobilization method for convenient biomolecule attachment in the context of immunosensors by inserting a “passivating” layer of PEI between the biomolecule and the surface. The primary criteria were the applicability of the surface modification to virtually any platform design with focus on microfluidics and the construction of scalable, functionally controllable biomolecular surface. We demonstrated the applicability of the disposable microfluidic chip to a wide range of immunoassay formats and diverse analytes in clinically relevant samples. Herein, the optimized cTnI assay utilized two capture mAbs 19C7/16A11 and one detection mAb MF4 to maximize the cardiac specificity and sensitivity while maintaining the highest signal possible. The LOD of 18.5 pg ml−1 and imprecision ≤12% CV was measured on the microtiter plate. The chip performance was assessed with two immobilization approaches. The direct method based on EDC/SNHS coupling yielded poor assay characteristics. Both PEI-based procedures were extremely powerful since they could immobilize large amounts of biomolecules while showing higher biological activity in comparison to the EDC/SNHS coupling. The adsorbed LPEI showed potential due to its simplicity and compatibility with various materials; however, failed to provide sufficient stability (approx. two days) against a high number of washing steps in comparison to immobilized BPEI (approx. six days). In contrast, the BPEI modified chip allowed reaching an LOD of 25 pg ml−1 with 10% CV, with other assay parameters comparable to the benchtop ELISA. The main merit of BPEI surface modification will strongly depend on its application; we have observed up to 60-times higher biomolecule loading/surface activity and approximately three-times lower nonspecific adsorption. Additionally, the stability of attached probes
[1] T. Porstmann, S.T. Kiessig, Enzyme immunoassay techniques an overview, J. Immunol. Methods 150 (1992) 5–21. [2] P. Tijssen, A. Adam, Enzyme-linked immunosorbent assays and developments in techniques using latex beads, Curr. Opin. Immunol. 3 (1991) 233–237. [3] C.D. Chin, T. Laksanasopin, Y.K. Cheung, D. Steinmiller, V. Linder, H. Parsa, et al., Microfluidics-based diagnostics of infectious diseases in the developing world, Nat. Med. 17 (2011) 1015–1019. [4] S.R. Narayanan, L.J. Crane, Affinity chromatography supports: a look at performance requirements, Trends Biotechnol. 8 (1990) 12–16. [5] P. Cutler, Affinity chromatography, in: Protein Purification Protocols, Springer, Totowa, New Jersey, 1996, pp. 157–168. [6] T.I. Yin, Y. Zhao, J. Horak, H. Bakirci, H.H. Liao, H.H. Tsai, et al., A microcantilever sensor chip based on contact angle analysis for a label-free troponin I immunoassay, Lab Chip 13 (2013) 834–842. [7] X. Fan, I.M. White, S.I. Shopova, H. Zhu, J.D. Suter, Y. Sun, Sensitive optical biosensors for unlabeled targets: a review, Anal. Chim. Acta 620 (2008) 8–26. [8] P. Esser, Activity of adsorbed antibodies, Appl. Note 11b (1997). [9] J.N. Herron, H.K. Wang, V. Janatova¡, J.D. Durtschi, D.A. Christensen, K.D. Caldwell, et al., Orientation and activity of immobilized antibodies, in: Biopolymers at Interfaces, second ed., 2003, pp. 115–164. [10] D. Kim, A.E. Herr, Protein immobilization techniques for microfluidic assays, Biomicrofluidics 7 (2013) 041501. [11] A. Subramanian, S.J. Kennel, P.I. Oden, K.B. Jacobson, J. Woodward, M.J. Doktycz, Comparison of techniques for enzyme immobilization on silicon supports, Enzyme Microb. Technol. 24 (1999) 26–34. [12] P. Jonkheijm, D. Weinrich, H. Schröder, C.M. Niemeyer, H. Waldmann, Chemical strategies for generating protein biochips, Angew. Chem. 47 (2008) 9618–9647. [13] W.D. Oliveira, W.G. Glasser, Hydrogels from polysaccharides. I. Cellulose beads for chromatographic support, J. Appl. Polym. Sci. 60 (1996) 63–73. [14] B. Johnsson, S. Lofas, G. Lindquist, A. Edstom, R.-M.M. Hillgren, A. Hansson, Comparison of methods for immobilization to carboxymethyl dextran sensor surfaces by analysis of the specific activity of monoclonal antibodies, J. Mol. Recognit. 8 (1995) 125–131. [15] R. Wong, H. Tse, Lateral Flow Immunoassay, Springer, 2009. [16] J. Horak, C. Dincer, H. Bakirci, G. Urban, A disposable dry film photoresist-based microcapillary immunosensor chip for rapid detection of Epstein-Barr virus infection, Sens. Actuators B: Chem. 191 (2014) 813–820. [17] P. Vicennati, A. Giuliano, G. Ortaggi, A. Masotti, Polyethylenimine in medicinal chemistry, Curr. Med. Chem. 15 (2008) 2826–2839. [18] R.R. Burgess, Use of polyethyleneimine in purification of DNA-binding proteins, Methods Enzymol. 208 (1991) 3–10. [19] M. Teramoto, H. Nishibue, K. Okuhara, H. Ogawa, H. Kozono, H. Matsuyama, et al., Effect of addition of polyethyleneimine on thermal stability and activity of glucose dehydrogenase, Appl. Microbiol. Biotechnol. 38 (1992) 203–208. [20] J. Yakovleva, R. Davidsson, M. Bengtsson, T. Laurell, J. Emneus, Microfluidic enzyme immunosensors with immobilised protein A and G using chemiluminescence detection, Biosens. Bioelectron. 19 (2003) 21–34.
J. Horak et al. / Sensors and Actuators B 209 (2015) 478–485 [21] Y. Bai, C.G. Koh, M. Boreman, Y.J. Juang, I.C. Tang, L.J. Lee, et al., Surface modification for enhancing antibody binding on polymer-based microfluidic device for enzyme-linked immunosorbent assay, Langmuir 22 (2006) 9458–9467. [22] M.M. Andersson, R. Hatti-Kaul, Protein stabilising effect of polyethyleneimine, J. Biotechnol. 72 (1999) 21–31. [23] M. Erol, H. Du, S. Sukhishvili, Control of specific attachment of proteins by adsorption of polymer layers, Langmuir 22 (2006) 11329–11336. [24] J. Horak, C. Dincer, H. Bakirci, G. Urban, Sensitive, rapid and quantitative detection of substance P in serum samples using an integrated microfluidic immunochip, Biosens. Bioelectron. 58 (2014) 186–192. [25] B. McDonnell, S. Hearty, P. Leonard, R. O’Kennedy, Cardiac biomarkers and the case for point-of-care testing, Clin. Biochem. 42 (2009) 549–561. [26] A.H.B. Wu, Cardiac Markers, Humana Press Inc., 2003. [27] C.S. Farah, F.C. Reinach, The troponin complex and regulation of muscle contraction, FASEB J. 9 (1995) 755–767. [28] S.J. Maynard, I.B.A. Menown, A.A.J. Adgey, Troponin T or troponin I as cardiac markers in ischaemic heart disease, Heart 83 (2000) 371–373. [29] J. Sarko, C.V. Pollack Jr., Cardiac troponins, J. Emerg. Med. 23 (2002) 57. [30] B. Cummins, M.L. Auckland, P. Cummins, Cardiac-specific troponin I radioimmunoassay in the diagnosis of acute myocardial infarction, Am. Heart J. 113 (1987) 1333–1344. [31] A. Katrukha, A. Bereznikova, V. Filatov, T. Esakova, Biochemical factors influencing measurement of cardiac troponin I in serum, Clin. Chem. Lab. Med. 37 (1999) 1091–1095. [32] T. Keller, T. Zeller, F. Ojeda, S. Tzikas, L. Lillpopp, C. Sinning, et al., Serial changes in highly sensitive troponin I assay and early diagnosis of myocardial infarction, J. Am. Med. Assoc. 306 (2011) 2684–2693. [33] J.P. Chapelle, M.C. Aldenhoff, L. Pierard, J. Gielen, Comparison of cardiac troponin I measurements on whole blood and plasma on the stratus CS analyzer and comparison with AxSYM, Clin. Chem. 46 (2000) 1864–1866. [34] E. Giannitsis, H.A. Katus, 99th percentile and analytical imprecision of troponin and creatine kinase-MB mass assays: an objective platform for comparison of assay performance, Clin. Chem. 49 (2003) 1248. [35] M. Panteghini, F. Pagani, K.T.J. Yeo, F.S. Apple, R.H. Christenson, F. Dati, et al., Evaluation of imprecision for cardiac troponin assays at low-range concentrations, Clin. Chem. 50 (2004) 327. [36] J.R. Tate, Troponin revisited 2008: assay performance, Clin. Chem. Lab. Med. 46 (2008) 1489–1500. [37] J. Wang, Electrochemical glucose biosensors, Chem. Rev. 108 (2008) 814–825. [38] G.T. Hermanson, Bioconjugate Techniques, second ed., Elsevier Science, 2008.
485
[39] I. Migneault, C. Dartiguenave, M.J. Bertrand, K.C. Waldron, Glutaraldehyde: behavior in aqueous solution, reaction with proteins, and application to enzyme crosslinking, BioTechniques 37 (2004) 790–806.
Biographies Josef Horak received his master’s degree in Biotechnology from the Brno University of Technology. This was followed by a one year scholarship at the Technical University of Lisbon, where he worked on cell based sensors. In 2013, he received his Ph.D. from the University of Freiburg, Germany, Department of Microsystems Engineering, where he is currently working as a group leader in the field of bioanalytical microsystems with a focus on immunosensors. Can Dincer received his master’s degree in Microsystems Engineering from the University of Freiburg in 2009, where he later worked as a scientific staff member at the Laboratory for Sensors. Since 2013, Can Dincer is head of the Nanosensors group in the field of immunosensors, microtechnology and sensor applications in medicine. Edvina Qelibari works as a student assistant at the Department of Microsystems Engineering at the University of Freiburg. Huseyin Bakirci studied Biochemistry at the University Louis Pasteur in France. In 2006, he obtained his Ph.D. in chemistry from the Department of Physical Chemistry, University of Basel. He continued at the same University as a postdoctoral researcher in the Department of Chemistry and Biochemistry and later on at the Division of Nephrology and Hypertension at the University Hospital of Bern in Switzerland. From 2007 till 2013, he was a group leader of bioanalytical microsystems at the Department of Microsystems Engineering, University of Freiburg. Gerald Urban studied technical physics at the TU Vienna, where he subsequently obtained his Ph.D. from the Department of Electrical Engineering. In 1985, he cofounded the company OSC in Cleveland, Ohio and in Vienna. He finished his post-doctoral studies in 1987 at the neurophysiological department in Münster, Germany. From 1990 till 2002 he was scientific director of the Ludwig Boltzmann Institute of Biomedical Microengineering in Vienna. He finished his habilitation in 1994 and became a full professor at the University of Freiburg in 1997. From 1999 to 2002 he was dean of the Faculty for Applied Science at the University of Freiburg.