Low-Cost and Rapid-Production Microfluidic Electrochemical Double

Sep 17, 2018 - Uncovering the Formation of Color Gradients for Glucose Colorimetric Assays on Microfluidic Paper-Based Analytical Devices by Mass ...
0 downloads 0 Views 954KB Size
Subscriber access provided by - Access paid by the | UCSB Libraries

Technical Note

Low-cost and rapid-production microfluidic electrochemical doublelayer capacitors for fast and sensitive breast cancer diagnosis Ricardo A.G. de Oliveira, Caroline Y. N. Nicoliche, Anielli M. Pasqualeti, Flavio Makoto Shimizu, Iris R. Ribeiro, Matias Eliseo Melendez, Andre L. Carvalho, Angelo Luiz Gobbi, Ronaldo C. Faria, and Renato Sousa Lima Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b02605 • Publication Date (Web): 17 Sep 2018 Downloaded from http://pubs.acs.org on September 18, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Low-cost and rapid-production microfluidic electrochemical double-layer capacitors for fast and sensitive breast cancer diagnosis Ricardo A. G. de Oliveira,†,# Caroline Y. N. Nicoliche,†,# Anielli M. Pasqualeti,†,# Flavio M. Shimizu,†,# Iris R. Ribeiro,†,‡ Matias E. Melendez,§ André L. Carvalho,§ Angelo L. Gobbi,† Ronaldo C. Faria,|| and Renato S. Lima*,†,‡ †

Laboratório Nacional de Nanotecnologia, Centro Nacional de Pesquisa em Energia e Materiais, Campinas, São Paulo 13083-970, Brasil Instituto de Química, Universidade Estadual de Campinas, Campinas, São Paulo 13083-970, Brasil § Centro de Pesquisa em Oncologia Molecular, Hospital de Câncer de Barretos, Barretos, São Paulo 14784-400, Brasil || Departamento de Química, Universidade Federal de São Carlos, São Carlos, São Paulo 13565-905, Brasil ‡

ABSTRACT: This technical note describes a new microfluidic sensor that combines low-cost (US$0.97) with rapid fabrication and user-friendly, fast, sensitive, and accurate quantification of breast cancer biomarker. The electrodes consisted of cost-effective bare stainless-steel capillaries, whose mass-production is already well-established. These capillaries were used as received, without any surface modification. Microfluidic chips containing electrical double-layer capillary capacitors (µEDLC) were obtained by a cleanroom-free prototyping that allows the fabrication of dozens to hundreds of chips in 1 h. This sensor provided the successful quantification of CA 15-3, a biomarker protein for breast cancer, in serum samples from cancer patients. Antibody-anchored magnetic beads were utilized for immunocapture of the marker and, then water was added to dilute the protein. Next, the CA 15-3 detection (< 2 min) was made without using redox probes, antibody on electrode (sandwich immunoassay), or signal amplification strategies. In addition, the capacitance tests eliminated external pumping systems and precise volumetric sampling steps, as well as presented low sample volume (5 µL) and high sensitivity using bare capillaries in a new design for double-layer capacitors. The achieved limit-of-detection (92.0 µU mL–1) is lower than the most methods reported in the literature for CA 15-3, which are based on nanostructured electrodes. The data shown in this technical note support the potential of the µEDLC towards breast cancer diagnosis even at early stages. We believe that accurate analyses using a simple sample pretreatment such as magnetic field-assisted immunocapture and cost-effective bare electrodes can be extended to quantify other cancer biomarkers and even biomolecules by changing the biorecognition element. KEYWORDS: impedance spectroscopy, magnetic field, immunocapture, CA 15-3 biomarker, early cancer diagnosis. In recent years, microfluidic devices integrating electrochemical sensors (µES) have been widely employed for cancer diagnosis targeting the quantification of biomarker proteins.1,2 In addition to the advantages of microfluidics such as low sample consumption, rapid analyses, and automation compatibility, the electrochemical techniques afford simple, multiplex, sensitive, and selective tests.3–7 While benchtop potentiostats undermine the accomplishment of in-situ analyses by nonspecialized users, handheld modules reported in the literature are cost-effective (approximately US$20.0) and can be interfaced to smartphones using wireless communication for point-ofcare (POC) applications.8,9 Despite of the previous advantages, both the chip fabrication and signal amplification strategies promoted through nanomaterials, enzymes, or DNA typically damage the cost, large-scale manufacturing compatibility, and user-friendliness of the µES,10 undermining their commercialization and high-volume tests for either personal clinical routine health checks (in home healthcare settings and in developing countries by first responders) or high-end biomedical diagnoses of serious diseases.11,12 In addition to the patterning of the microfluidic channels, the µES fabrication includes the incorporation of embedded electrodes. In general, the µES used for cancer biomarkers are obtained from conventional high-cost and complex cleanroomassisted methods such as photolithography, soft-lithography, thin film deposition, and wet etching.13–16 Cost-effective alternatives to engrave the channels include xurography of doublesided adhesives,17 3D printing on polymers,18 and wax printing on paper,7,19–21 whereas the electrodes are commonly attained through screen-printing technology.20–23 While the other methods show poor to moderate throughput, the wax printing is expected to be transferred to industrial-scale manufacturing

and screen-printed electrodes are commercially available in distinct designs. However, just like low-cost methods such as drop casting and inkjet printing, the screen printing can leave residues on the substrate surface after solvent evaporation.24 The latter approach further produces a large amount of wastage material.25 One should also emphasize that thick-film electrode fabrication methods free of vacuum deposition and cleanroom facilities (screen and inkjet printing, drop casting, and direct writing utilizing pencil or ball pen) are only able of manufacturing polarized quasi-reference electrodes, which do not provide stable and well-defined reference potentials in different faradaic electrochemical assays.26 In practice, reliable thick-film reference electrodes are achieved by surface treatments such as chemical chlorination25 and electrodeposition26 that form layers of Ag/AgCl. With regard to the signal amplification methods, they commonly require costly reagents as well as laborious and time-consuming steps of surface treatment for both capacitive and faradaic sensing devices.10 All the aforesaid shortages hinder the translation of the sensors into practical use. For instance, no commercial microfluidic paper-based analytical device (µPAD) is still capable of detecting cancer marker proteins in clinically-relevant samples in spite of the mass-production compatibility of the wax printing fabrication.7 Herein we address a sensor that combines low-cost with both rapid fabrication and simple, fast, and sensitive determination of breast cancer biomarker. Microfluidic chips containing electrochemical double-layer capillary capacitors (µEDLC) in polydimethylsiloxane (PDMS) substrate were obtained by a bondless and cleanroom-free prototyping that basically needs a laboratory oven and allows the fabrication of dozens to hundreds of chips in 1 h.27–31 The electrodes consist1

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ed of low-cost stainless-steel capillaries, whose large-scale production is well-established. These electrodes were reversibly introduced into the channels and utilized as received, without any surface chemical modification. This feature creates reusable electrodes as well as avoids steps to irreversibly attach electrodes in the microfluidic chip, representing relevant advantages even over methods such as screen-printing towards mass-production and, then commercialization and high-volume applications. The sensor provided user-friendly, rapid, and accurate breast cancer diagnosis in serum samples from human patients by using antibody-anchored magnetic beads (MBs-Ab). Such MBs-Ab have been successfully used to capture cancer biomarkers, leading to accurate determinations.32–38 After the immunocapture of the protein by the antibody-anchored MBs, this conjugate is separated from sample applying a magnetic field. Next, electrolyte is added in substitution to the biological fluid to dilute the tumor marker. This protocol ensures selective analyses by eliminating interferents present in the samples, a key routine in our case because of the universal character of the µEDLC. When coupled with faradaic detection, such methods generally rely on a sandwich immunoassay utilizing i) primary antibodies (Ab1) covalently linked to the working electrode and ii) active probe and secondary antibodies (Ab2) on the MBs. The necessity to further chemically modify the MBs-Ab2 with redox mediating probes and the electrodes with both Ab1 and signal amplification species damages the analysis time other than the cost and userfriendliness of the method because of the incubation time that is required for binding among the immunocaptured protein and Ab1.17,39 In contrast, here we performed the capacitance readout for protein detection without using redox probes, signal amplification, or antibodies on electrode, Ab1. One advantage of capacitive sensors on faradaic methods that is relevant for POC tests should also be emphasized. This merit lies in the absence of non-polarized reference electrodes, thus leading to a simpler instrumental setup and avoiding hurdles related to the cell potential stability.29 Another characteristic that contributes to µEDLC user-friendliness is the accomplishment of the measurements at stationary medium free of precise volumetric sampling steps and pumping systems. In practice, plastic syringes were used to clean the electrodes and channels as well as to change the samples. Such simplification is mandatory for POC assays. As example, the requirement of powered pumping systems to hand the liquids is considered to be a key hurdle in the market entry of conventional micro total analysis systems.7 The capacitance tests further demonstrated low sample volume (5 µL) and high sensitivity. Large surface areas are essential to raise the capacitance and, then the performance of supercapacitors as energy storage systems.40–44 In addition, electrodes presenting tubular structure ensures high ion-accessible sensing areas and fast mass transport of the ions, further increasing the values of capacitance as verified in nanostructured supercapacitors.42–44 Here, both these characteristics of the electrodes (high surface area and tubular structure) led to sensitive analyses even using bare metals. This sensitivity can be further improved by simply raising the number of capillaries acting as capacitors in parallel of the µEDLC. The sensor was applied in the detection of the carbohydrate antigen 15-3

Page 2 of 11

(CA 15-3) in serum samples from cancer patients. CA 15-3 glycoprotein (400 kDa) is a crucial marker for breast carcinoma that is widely utilized for early diagnosis as well as for progression and recurrence diagnosis.17

EXPERIMENTAL SECTION

 Chemicals and samples. Bovine serum albumin (BSA), 1(3-(dimethylamino) propyl)-3-ethylcarbodiimide (EDC), 4-(2hydroxyethyl)piperazine-1-ethanesulfonic acid (Hepes), 2-(Nmorpholino) ethanesulfonic acid (MES), and Tween-20 (T20) were purchased from Sigma-Aldrich (St. Louis, MO). The MBs (Dynabeads® MyOne™ Carboxylic Acid) were acquired from Thermo Fisher (Waltham, USA). Polyclonal Antibodies (Rabbit, AB62687) and Human Cancer antigen 15-3 fulllength protein (CA 15-3, AB182029) were purchased from ABCAM Lab. Ltd (Bristol, United Kingdom). The solutions were prepared in distilled water (Milli-Q, Millipore Corp., Bedford, MA), which was obtained with resistivity of 18 MΩ cm. Four human serum samples were supplied by Barretos Cancer Hospital (Barretos, Brazil). These samples were separated in aliquots of 5 µL and maintained at -80 °C. The CA 153 concentration in these samples was quantified through electrochemiluminescence (ECL) analyzer model Cobas 601 and ElecsysCA15-3 immunoassay kit, both from Roche Diagnostics (Indianapolis, IN). The use of serum samples was approved by the Research Ethics Committees of both Barretos Cancer Hospital and Federal University of São Carlos (São Carlos, Brazil). Device prototyping. The fabrication of the chips was based on previous works reported by our group.27–31 Microfluidic devices of a single piece of PDMS featured four to eight parallel channels that were produced through a bondless, low-cost, and cleanroom-free technique of sequential steps of polymerization and scaffold removal (PSR). Briefly, a scaffold of the channel is shaped and then PDMS is poured. Finally, the PDMS is cured and the scaffold is manually removed. To prepare PDMS, monomer and curing agent were mixed at 10:1 w/w and, then degassed under vacuum for 20 min. The channels were achieved using stainless-steel capillaries with diameter of 700 µm as scaffolds that were mounted on a piece of aluminum fixed on a glass slide as illustrated in Figure S1A (Supporting Information). The PDMS cure was initiated at 50 °C in a conventional laboratory oven (Blue M, Blue Island, IL) and, then the temperature was raised to 65 °C. This cure was conducted for 1 h. Using two pieces of 34 mm × 56 mm × 22 mm, 12 chips could be obtained in approximately 1 h. Therefore, this technique could be easily transferred to large-scale fabrication by simply increasing the number of pieces with polymer-coated scaffold in a single thermal stage. PSR is further a green prototyping process by eliminating both the use of solvent and the generation of any waste. Electrodes. The electrodes consisted of stainless-steel capillaries (Treficap, São Paulo, Brazil) with 800 µm in diameter and i.d. of 545 µm. Steel capillary pairs were manually placed in the PDMS channels as shown in Figure S1A. This process was conducted with the aid of magnifying glass to obtain electrode gaps (distances between the capacitor ‘plates’) of roughly 200 µm as shown in Figure 1A. The ends of the capillaries facing to each other in the channels were previously 2

ACS Paragon Plus Environment

Page 3 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

polished using P600 aluminum oxide and 1200 silicon carbide sandpapers. The electrodes were short-circuited using copper pieces for producing an association of capacitors in parallel. Additionally, the capillaries were connected to each other utilizing Tygon tubes (0.7 mm i.d.) to allow the filling of all the fluidic circuit by the samples from one single plastic syringe, which was utilized to clean the electrodes and channels between the capacitive assays with water and ethanol as well as to change the samples. Images of the electrodes were obtained by scanning electron microscopy (SEM, FEI Quanta 650 Thermo Scientific, Ashford, UK), stereoscopy (Leica M125, Wetzlar, Germany), and atomic force

3

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 11

Figure 1. Device and electrodes. (A) Chip filled with red dye and short-circuited electrodes (I,II) and stereoscopy images of the cross-section of one steel capillary in the channel (III) and of the gap between capillaries that produce one capacitor (IV). (B) SEM, (C) 3D AFM surface topographies, and (D) XPS spectrum of the inner walls of a capillary. In (C), II and III were extracted from specific regions of I as shown in Figure S1B.

microscopy (AFM, NX10 Park Systems, Santa Clara, CA). The AFM images were acquired in areas of 20 µm × 20 µm in the tapping mode at 0.3 Hz using a FMR tip of Nanosensors (Neuchâtel, Switzerland). The stereoscopy images were valuable to measure the channel and electrode diameters using the software LAS Core V3.8. In addition, X-ray photoelectron spectroscopy (XPS) analyses of the inner walls of the electrodes were made in the Thermo Fisher Scientific K-Alpha spectrometer. Analyses. Impedance spectroscopy analyses were performed in Metrohm Autolab AG PGSTAT302N (Herisau, Switzerland) and PalmSens4 (PalmSens, Houten, Netherlands) potentiostats. The latter one was controlled by a smartphone. All these measurements were performed at stationary media applying a voltage of 50 mV ac, integration time of 2 s, and frequencies from 1 to 106 Hz at room temperature. The tests (in triplicate) lasted less than 2 min. The capacitances recorded at a same frequency were selected for constructing the analytical curves. Modification of the MBs with polyclonal antibodies (MBs-Ab). According to a procedure reported in the literature for MB modification with polyclonal Ab,17 200 µL of MBs were transferred to a microtube containing 1 mL of MES buffer (15 mmol L–1, pH 6.0) and vortexed for 30 s. Next, these MBs were magnetically separated using a magnetic rack for 2 min and the supernatant was removed. This step was repeated 3 times. Following, the MBs were dispersed in 1 mL of MES containing EDC (10 mg mL–1) for activation of carboxylic groups, incubated under slow stirring for 30 min at room temperature, and washed three times with MES. After, 1 mL of Ab diluted in MES buffer (10 µg mL–1) was added and incubated under slow stirring overnight. Using the magnetic rack, the MBs-Ab was separated and, then washed with 1 mL

of PBS-T20 (twice). These conjugates were dispersed in 1 mL of glycine (1 mol L–1, pH 8.0) and vortexed for 30 min. The resulting mixture was washed with MES three times and dispersed in 400 µL of PBS-T20 0.1% v/v to BSA to block nonbioconjugated active sites of the MBs. Immunocapture. The volume of the magnetic particles modified with polyclonal antibody added in the immunocapturing step and the capture time were employed in agreement with optimization studies reported in the literature.17 To capture CA 15-3 standards in order to get analytical curves, 20 µL of MBs-Ab were added to 340 µL of PBS buffer (pH 7.4) with CA 15-3 at the desired concentration and incubated for 70 min at room temperature. Then, the MB-Ab-CA 15-3 conjugates were magnetically separated and washed with 400 µL of distilled water three times. Succeeding, such conjugates were simply diluted in 125 µL of distilled water and injected into the µEDLC to make impedance tests. These MBs-Ab were also used for application to the real samples. In this case, 5 µL of each serum sample were first diluted in 500 µL of PBS buffer 0.5% v/v to BSA (pH 7.4). After homogenization in vortex for 30 s, 10 µL of this solution were diluted in 1 mL of PBS and 340 µL were added in a microtube to make the immunocapture after adding 20 µL of MBs-Ab as described above.

RESULTS AND DISCUSSION

 final µEDLC with integrated electrodes is presented in The Figure 1A, as well as stereoscopy images of the cross-section of one stainless-steel capillary and of the gap among two capillaries. This gap (approximately 200 µm) defined the distance between the capacitor electrodes. SEM and AFM images 4

ACS Paragon Plus Environment

Page 5 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

demonstrated that the inner walls of the capillaries have a highly rough surface as displayed in Figure 1B and 1C, respectively. This increased surface area is mandatory to enhance the capacitance of the sensor as aforesaid. From XPS survey spectrum in Figure 1D, the main peaks are assigned to carbon (C) 1s, oxygen (O) 1s, chromium (Cr) 2p and 3p, and iron (Fe) 2p3/2 and 3p as highlighted. These data are in accordance with the surface of stainless-steel. All the relative composition of the capillary is shown in Figure S2. The µEDLC capacitors can be assumed as ideally polarizable electrodes because of the absence of faradaic reactions. In fact, this sensor relies on the charging of the solution in the electrical double layer (EDL), with the reversible adsorption of both anions and cations of the electrolyte onto the electrode surfaces.40 In contrast to solid capacitors, the charge and, then the absolute phase potential exhibits a transient distribution in the EDL because of the existence of a surface tension in the solution, which expresses the energy required to increase the EDL area as a unit value.45 This tension diminishes with the potential and concentration of ions in the solution, leading to a consequent narrowing of the EDL. In this regard, the charge and usually the capacitance (Cd) of the metal/solution interface are proportional to the potential and concentration. Otherwise, the capacitance of electrical analogues remains constant with these two variables. The variation profile of the EDL depends also on the occurrence of phenomena such as nonspecific and specific adsorptions, coupled reactions, and charge transfer in pseudocapacitors.40,46 The potential transient distribution and electrical model of a charged metal/solution interface assuming an ideally polarizable electrode are demonstrated in Figure 2A. In this model, the faradaic impedance is neglected and the capacitive process can be represented by simple linear circuit elements like Cd and solution resistance (Rs).45 Nonetheless, one should underline that this capacitor does not show equivalent behavior to its electrical analogous as aforementioned. The effect of the number of capacitors, electrode length, and medium salinity over the analytical performance of the µEDLC was evaluated through measurements to aqueous solutions of a small ionic electrolyte, KCl. By considering the absence of coupled and charge-transfer reactions, the capacitance is expected to increase with the KCl content as it was indeed noted in Figure 2B that shows capacitance vs. frequency spectra changing the KCl concentration from 10.0 to 500.0 ppm. In this case, the chip was composed of four capacitors with 20-mm length capillaries. The volume of each KCl solution used for the measurements (n = 4) was lower than 1.0 mL. According to the analytical curves (capacitances at 103 Hz since this frequency led to the best linearity, with correlation coefficients larger than 0.99) and limits-of-detection (LOD) in Figure 2C, the increase in the number of parallel capacitor was proven to be an effective mode to improve the sensibility and detectability of the method. This signal amplification mode is simpler than the surface modification-based methods that involve complex and laborious steps and high-cost chemicals.10 In addition to the low-cost of the steel electrodes, the µEDLC experimental protocol is independent on the number of parallel capacitors inserted into the device. Chips with 1, 4, and 8 capacitors based on 20-mm length capillaries exhibited increasing analytical sensitivities of 10.4, 42.3, and 63.0 10–12 ppm nF–1, respectively. The lower LOD (calculated as 3S/m

where S is the blank standard deviation for n = 4 and m is the slope of the analytical curve) attained with 8 capacitors was 80.5 ppb. The length of the electrodes also changes the µEDLC sensitivity according to data for a chip with 4 capacitors and 10 mm-length capillaries. This assembly led to a somewhat

Figure 2. Electrical model and analytical performance of the µEDLC. (A) Model and electrical circuit of the metal/liquid interface for an ideally polarizable electrode (IPE) at slightly negative potential assuming specific adsorptions of water (gray) and anions (yellow) and nonspecific adsorption of cations (green). (B) Spectra for KCl standard using the chip with 4 capacitors and 20-mm length capillaries. The capacitances raised with KCl contents as indicated by the red arrow and the symbol * highlights the frequency chosen for the capacitances in the analytical curve. These analyses were made using the benchtop potentiostat. (C) Analytical curves (I) and LODs (II) for KCl. In these curves, ‘n’ is the number of parallel capacitors in the µEDLC for 20-mm (n*) and 10-mm (n**) length capillaries. In the LOD plots, A represents the µEDLC with 1 capacitor and 20-mm length capillaries, B and C express the use of 4 capacitors with 10mm and 20-mm length capillaries, respectively, and D is related to the chip containing 8 capacitors and 20-mm length capillaries. The insets show the circuit of parallel capacitors (I) and photos of the devices that were employed as defined by B (10-mm) and D (20-mm length capillaries) (II).

poorer analytical sensitivity (31.0 10–12 ppm nF–1) and LOD (129.9 ppb) compared with the chip containing 20-mm length capillaries due to the decrease in sensing area. The µEDLC sensitivity is further expected to increase with a reduction in the electrode gap. However, distances lower than 200 µm were not tested because of the creation of air bubbles in this region, leading to unreproducible values of capacitance. The µEDLC showed satisfactory reproducibility for measurements to 100.0 ppm KCl in five distinct chips. In one of these cases, the position of the electrical contacts with the electrodes was also changed. The global average capacitance (103 Hz) was 6.5 nF with a confidence interval of only 5.3 pF 5

ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(α = 0.05 and n = 30), confirming that the visual process to achieve electrode gaps of approximately 200 µm as well as the device fabrication and electrodes did not undermine the reproducibility of the capacitance recording assays. As disadvantage, the µEDLC shows poor sensitivity for high-salinity media. Using 4 capacitors with 20-mm length capillaries, the analytical sensitivity decreased from 42.3 10–12 to 27.0 10–12 ppm nF–1, whereas the LOD increased from 119.9 to 187.1 ppb when KCl was prepared in saline aqueous solution of NaCl 100.0 ppm rather than distilled water. In addition, the previous values were changed to 15.4 10–12 ppm nF–1 and 329.0 ppb when KCl was prepared in NaCl 1,000.0 ppm. This outcome is associated with a remarkable decrease in the transport number of the analyte. In this case, the excess electrolyte practically assumes all the contribution of the migration to the total current flowing in the external circuit of the sensor. Therefore, the analyte is only transported by diffusion, which is poor in electrochemical double-layer capacitors owing to the application of ac voltage and absence of formation or consumption of species in the EDL. Both these aspects contribute to reduce the diffusion flux by decreasing the gradient between the analyte amount in the solution bulk and EDL.45 In the µEDLC, this gradient is created only by the ion polarization. The challenges facing the use of the µEDLC to quantify cancer biomarkers in human serum samples are the poor sensitivity of the sensor in high-salinity conditions and the universal character of capacitive methods, thus undermining the selectivity of the determinations. Herein the magnetic immunocapture of the protein solved both the previous limitations by selectively capturing the target protein, removing the interferents present in the samples, and providing the possibility to subsequently replace the sample by low-conductivity media like distilled water. The routine of this sample pretreatment relies on four main steps as demonstrated in Figure 3A, namely, addition of MBs-Ab in the sample, stirring to selectively capture the CA 15-3 protein, magnetic separation of the MB-Ab-CA 15-3 conjugates followed by washing and dilution with distillate water, and manual insertion of the sample into the device for capacitance detection simply using a plastic syringe. The capacitance vs. frequency spectra and resulting analytical curve to CA 15-3 standards are shown in Figure 3B. In this case, we used a chip with four capacitors and 20-mm length capillaries. While other frequencies led to analytical curves with linear fitting, the capacitances at 21.5 103 Hz were taken up by generating the best accuracy in the application to samples of human serum. The MB-Ab concentration was the same in all the assays to get the analytical curve. The increase in the capacitance with the building up of CA 15-3 anchored on the MBs-Ab is likely due to the covalently bonded oxygenated functional groups over the protein surfaces, which increase the negative overall surface charges of the conjugates.47 The correlation coefficient in analytical curve was 0.99 and the limit of linearity and LOD were 5.0 mU mL–1 and 92.0 µU mL–1, respectively. The latter parameter (LOD) was not improved by raising the number of parallel capacitors since it is already below the threshold value for CA 15-3, 25.0 U mL–1.48 Anyway, this low limit-of-detection is advantageous by i) providing early diagnosis, ii) decreasing the required volume of

Page 6 of 11

serum from cancer patient, which is diluted in buffer (herein, e.g., only 5 µL were needed for analyses in triplicate), and iii) generating the possibility to decrease the Ab/CA 15-3 binding time. Herein, this time was 70 min to attain an optimum sensitivity. According to results of the literature for oxidation of hydroquinone catalyzed by horseradish peroxidase (HRP) that was anchored onto MB-Ab2 conjugates in sandwich immunoassays selective to CA 15-3,17 a capture time of only 10 min led to a decrease in the peak current of 38% compared with

Figure 3. Measurements for CA 15-3 standards. (A) Sample pretreatment based on four steps: addition of MBs-Ab in the sample (I), stirring for biomarker immunocapture (II), magnetic separation of the MB-Ab-CA 15-3 conjugates followed by subsequent washing and dilution with distillate water (III), and pumping into the chip to measure the capacitance (IV). (B) Capacitance spectra (I) and analytical curves for CA 15-3 standards using 4 capacitors and 20-mm length capillaries (II). These analyses were made using the benchtop potentiostat. The capacitances in the spectra increased with the CA 15-3 concentration as indicated by the red arrow, whereas the symbol * highlights the frequency chosen for the capacitances of the analytical curve.

the data recorded after a capture time of 70 min. In our case, the increase in the number of parallel capacitors offers a simple way to provide satisfactory detectability even when using an analyte capture time as low as 10 min. The reduction in analysis time is an important goal towards POC applications. In addition, this improvement on detectability may be mandatory to detect other biomarkers such as prostate specific antigen (PSA), carcinoembryonic antigen (CEA), and alphafetoprotein (AFP), which are found in the sera at much lower concentrations ranging from 4 up to 10 ng mL–1.48,49 In relation to the impedimetric techniques to quantify cancer biomarkers without the use of redox probes, the LOD obtained 6

ACS Paragon Plus Environment

Page 7 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

herein is notably lower than values reported for CA19-9 using multilayer film architectures (690.0 mU mL–1)50 and interdigitated electrodes modified with carbon nanotubes (350.0 mU mL–1).51 Concerning the determination of CA 15-3, our value is also lower than the most methods described in the literature that include methods such as voltammetry, ECL, surface plasmon resonance, optofluidic ring resonator, and fluorescence in both batch or microfluidic systems such as µPDAs (Table S1). These methods show values of LOD from 4.0 mU mL–1 up to 15.0 U mL–1. Exceptionally, an amperometric device proposed by our group based on magnetic capture and sandwich immunoassays (HRP probe as aforementioned) showed the best detectability,17 with a LOD of 6.0 µU mL–1. As advantages on the latter chip, nonetheless, the µEDLC eliminates modifications of the MBs-Ab with redox probe and of the electrode with primary Ab (we only modified the MBs with Ab, which is the biological recognition element). Each one of these routines lasts overnight and involves laborious steps and high-cost chemicals and. Additionally, the capacitance detection (