Developing enhanced magnetoimmunosensors based on low-cost screen-printed electrode devices

SPE: definition and characteristics

SPE are produced by screen printing, which consists of stamping the electrodes layer-by-layer using conductive inks and mesh screens (Figure 2). Hence, SPE are fabricated by a low-cost and mass-production-compatible technology that allows the fabrication of devices significantly cheaper than most classical electrodes (Tudorache and Bala 2007, Metters et al. 2011, Couto et al. 2016). Printing is done onto a physical substrate (a plate or sheet) made of insulating material (e.g. a polymer, plastic, ceramic, alumina, glass, paper). A number of mesh screens, usually three to four, are then sequentially employed, each one displaying a series of perforations that delimit the geometry of the different electrodes/components. The inks used, which will compose the electrodes once printed and provide them their functional characteristics, are made of a mixture of a conductive material (such as graphite or metal powder), solvents (terpineol, ethyl cellosolve, cyclohexanone, ethylene glycol), a polymeric binder (glass powder, resins, cellulose acetate), and additives. When a screen is placed onto the substrate and the appropriate ink is added and squeezed over the screen, the ink penetrates through the perforations and is printed onto the solid surface. After ink drying and curing in an oven, the next screen and ink are added to the process. The whole procedure includes printing first silver tracks and contacts, followed by stamping the working, reference, and counter electrodes, and finishes with the addition of an insulating layer that delimits the different electrodes and the electric contacts. The different layers produced have thicknesses in the range of 20 to 100 μm and are thicker than those obtained by other fabrication methodologies (such as microfabrication by standard photolithography). For this reason, screen printing and SPE are referred to as “thick-film technology” and “thick-film electrodes” (Metters et al. 2011).

Figure 2:

Fabrication of SPE.

SPE are produced using conductive pastes/inks and mesh screens. For this, a conductive paste is placed onto the mesh screen and is squeezed over it (A). As a result, the ink passes through the pores in the screen and is printed onto the substrate below it (B). (C) An example of SPE of different materials.

The most widely used material for producing the working and counter electrodes is carbon/graphite, although a number of providers offer also SPE made of gold, silver, and platinum, among others (Table 1). Numerous authors have reported on the production of SPE displaying nanostructured working electrodes, in which a carbon working electrode [carbon SPE (C-SPE)] has been modified with graphene, carbon nanotubes (CNT), metal nanoparticles, etc. (Devi et al. 2015, Arduini et al. 2016), and the first nanostructured SPE have been already commercialized (Table 1). On the contrary, reference electrodes are usually produced with a silver or Ag/AgCl paste.

Compared to classical bulk electrodes, SPE display a number of advantages. For instance, SPE address the issue of cost-effectiveness because they can be mass produced and are easy and cheap to fabricate. The low variability between devices claimed by authors and producers, often in the order of 4% to 14% (Fanjul-Bolado et al. 2007, García-González et al. 2008, Vidal et al. 2012, Maesa et al. 2013, Afonso et al. 2016), allows also highly reproducible and sensitive electrochemical detection of bioassays. Some commercially available SPE can be used straightforward, not requiring any pretreatment, such as electrode polishing or electrochemical preactivation. Because using low-cost disposable SPE prevents electrode surface poisoning and cross-contamination between measurements and samples, SPE circumvent also some of the common problems associated with the use of classical solid electrodes, such as memory effects and the need to carry out tedious and time-consuming cleaning processes before and between experiments. SPE are very convenient to use because they might integrate all the electrodes of the electrochemical cell in a single portable device, allowing the study of relatively small sample volumes (e.g. a drop of few microliters) and producing minimal waste and disposal of reagents. Accordingly, SPE provide an inexpensive disposable platform for the rapid and accurate detection of Ag. In addition, the possibility to produce the electrodes with commercially available inks of different composition, and/or modify them with different materials and nanocomponents, makes it possible to produce highly specific and finely calibrated electrodes for the detection of any specific target analyte (Metters et al. 2011). The recent technological advances have lead to the development of novel and highly sensitive SPE devices, including SPE produced in flexible and stretchable materials for the production of wearable sensors, nanostructured and chemically modified devices for enhanced detection of selected electroactive molecules, and the production of screen-printed microelectrodes and microelectrode arrays for ultrasensitive detection (Tudorache and Bala 2007, Metters et al. 2011, Thiyagarajan et al. 2014).

It has been repeatedly discussed whether SPE display poorer performance than other devices, such as glassy carbon electrodes (GCE), which are appreciated in the field for their low electrical resistivity, extreme resistance to chemical attack, impermeability to gases, and wide potential range (the widest among carbon electrodes). For instance, Grennan et al. demonstrated that, when detecting ferrocyanide by cyclic voltammetry (CV), C-SPE displayed larger peak-to-peak separation (100–120 mV) than GCE (66–76 mV) as well as heterogeneous electron-transfer rate constants five times lower (Grennan et al. 2001). Barallat et al. also observed that the chronoamperometric detection of peroxidase enzymes generated higher signals, lower LOD, and wider linear range at Pt thin-film electrodes with a stamped Ag/AgCl pseudoreference than at C-SPE and single-walled CNT (SWCNT)-SPE (Barallat et al. 2013). Such differences were attributed to the presence of a polymeric binder in the carbon ink, which may affect negatively the electrode performance. However, it has been also shown that the performance of some SPE improves, and can become even comparable to that of GCEs, when the inks are cured at higher temperatures (above 100–200°C; Wang et al. 1996, Grennan et al. 2001). Similar observations have been made in the case of electrode pretreatment. For instance, improved performance of C-SPE has been reported after anodization at 2.0 V versus Ag/AgCl for a suitable time in phosphate-buffered saline (PBS), H₂SO₄, or NaOH (Thiyagarajan et al. 2014). Besides, the composition of the inks and physical substrates used for screen printing can strongly influence the analytical performance of the resulting SPE, and several authors have observed important differences between SPE from different providers (Fanjul-Bolado et al. 2008, García-González et al. 2008). An example of successful implementation of SPE was reported by Maesa et al., who could detect skatole and indole, two porcine hormones, using disposable C-SPE with results comparable to those of a GCE, which avoided having to perform extensive electrode polishing, washing, or activation before or between measurements (Maesa et al. 2013).

Production of classical electrochemical immunosensors

The production of classical immunosensors is based on the incorporation of Ab to the working electrode surface by a variety of methods (Jung et al. 2008, Makaraviciute and Ramanaviciene 2013, Trilling et al. 2013). Random physisorption stands as the simplest one, with variable performance efficiency depending on the characteristics of the Ab, electrode, and target molecule (Laczka et al. 2008a). Some examples of alternative protocols are Ab entrapment via the (electro)polymerization of a suitable polymeric matrix, layer-by-layer deposition, sol-gel chemistry modification, and chemical cross-binding to reactive sites on the electrode surface. For this later strategy, many authors have exploited reactive groups, such as COOH, which could be incorporated by casting on the electrode carboxylated CNT, created by anodization of carbon materials, or introduced by monolayer self-assembly of carboxylated thiols on gold materials (Baldrich et al. 2008a,b). Diazonium chemistry has been shown also useful, especially, but not exclusively, on carbon materials (Chung et al. 2012).

To improve the sensitivity, selectivity, stability, and reproducibility of the immunosensor, Ab-coated SPE are often modified with additional components. For example, a number of reagents can be added to (1) improve Ab immobilization [e.g. chemical cross-linkers such as glutaraldehyde or 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDC) and N-hydroxysulfosuccinimide (NHS)], (2) stabilize the immunolayer (e.g. cellulose acetate and Nafion), (3) prevent the nonspecific binding of nontarget components [e.g. proteins such as bovine serum albumin (BSA), casein and gelatine; polymers such as dextran and polyethylene glycol (PEG)], or (4) improve the electrochemical detection of enzyme label activity in sandwich assay formats (e.g. redox mediators such as ferrocene derivatives, Meldola blue, Prussian blue, cobalt phthalocyanine, or (poly)thionine; enzyme cofactors such as NADH and PQQ; Tudorache and Bala 2007).

According to the transducer used and the measurement set-up, the signal measured by the sensor can be a current flow (i.e. amperometric sensors), change in potential or charge accumulation (i.e. potentiometric sensors), or alteration in the conductive properties of the medium between the electrodes or across the electrode-solution interface (i.e. conductometric and capacitive sensors; Figure 3). In all the cases, SPE electrochemical immunosensors should optimally detect target binding in a straightforward manner in a label-less assay format (Luo and Davis 2013). Among the available measurement techniques, electrochemical impedance spectroscopy (EIS) is the most widely used for the straightforward detection of the immuno-binding event, although often in the presence of redox probes. In this context, it is well known that Ag binding to the sensor surface contributes to change its behavior and properties (Daniels and Pourmand 2007, Alonso-Lomillo et al. 2010, Prodromidis 2010). For instance, the formation of the Ab-Ag complex on the surface of the electrode produces an increase in the dielectric layer thickness, causing changes in capacitance and electron transfer rate proportional to the size and concentration of the bound biomolecules. Impedimetric sensors display low detection limits and short and simple assay formats, but also very complex data analysis and signal interference by many external factors (such as cable length or electrical noise).

Figure 3:

Different strategies for the detection of electrochemical immunosensors.

(Left) In electrochemical immunosensing, signal generation can be accomplished by different means. From top to bottom, direct detection of the immunocapture event, detection using enzyme labels such as HRP and AP, detection using metal nanoparticles as enzyme mimetics, and detection by acidic dissolution of QD. (Right) Detection can be performed by different techniques, such as (from top to bottom) EIS, CV, DPV/SWV, and amperometry.

Alternatively, SPE immunodetection can be accomplished in sandwich assay formats using enzyme labels. In these cases, the amperometric monitoring of the enzymatic reaction is the method most usually carried out, which reportedly provides high sensitivity, although examples using differential pulse voltammetry (DPV), square wave voltammetry (SWV), and potentiometry have also been described (Mistry et al. 2014). For this, appropriate enzyme substrates have to be used, searching for those that generate minimal background currents on their own but that, in the presence of the enzyme label, release an electroactive reaction product. This reaction product is then oxidized or reduced at the electrode surface and the current intensity measured is directly proportional to the concentration of analyte in the sample. On the contrary, the implementation of nonenzymatic nanolabels, such as metal nanoparticles and quantum dots (QD), is mostly followed by the nanoprobe acidic dissolution and anodic stripping voltammetry (ASV) detection of the released ions (Ding et al. 2013, Huang and Zhu 2013, Pei et al. 2013).

Limitations of classical SPE electrochemical immunosensors

Examples of sensitive immunosensors using SPE have been developed for the detection of various biomarkers, including Plasmodium falciparum, histidine rich protein-2, carcinoembryogenic antigen, α-fetoprotein, cancer antigen 125, and aflatoxin B1, many of them displaying LOD ranging from few picograms/milliliter to tens of nanograms/milliliter (Tudorache and Bala 2007, Mistry et al. 2014). However, classical immunosensors display a number of limitations. First, the bioengineering of the electrode surface for its usage both as a solid phase and an electrochemical transducer results in the partial shielding of the surface, which in turn can cause hindrance of the electron transfer and result in reduced electrochemical signals compared to the use of an unmodified/bare electrode (Baldrich et al. 2009). Second, immunosensor development involves the optimization of numerous steps, including various incubations and washings. Although not usually discussed by the authors, such long procedures and extended exposition to aqueous medium might have a deleterious effect on SPE integrity and reproducibility between individual devices. Third, because electrodes have to be handled and submitted to detection one by one, immunosensor optimization entails also tedious and time-consuming work. Finally, biosensor reusability is restricted due to the permanent immobilization of the Ab or Ag molecules on the electrode surface and the limited success of the Ab regeneration protocols reported to date (Solé et al. 2001, Xu and Wang 2012, Goode et al. 2015). For these reasons, attempts to separate the surfaces used for immunocapture and electrochemical detection have been reported both using SPE and other types of devices (Laczka et al. 2011a,b). The utilization of MP stands as one of the most versatile and promising among the different strategies available.

Advantages of using MP for target immunobinding

The utilization of MP for the magnetic separation and concentration of target molecules is probably one of the most versatile separation processes in biotechnology and analytical chemistry. A good proof of it is that nowadays numerous providers sell MP and magnetic nanoparticles (MNP) of different types and properties (Table 2).

MP have been successfully employed to bind a variety of molecular targets (cells, viruses, proteins, nucleic acids, and drugs) from different types of matrices, including crude samples (Hsing et al. 2007, Borlido et al. 2013). In this context, MP provide unique advantages compared to other separation techniques (e.g. filtration, centrifugation, or dielectrophoresis) such as a fast, efficient, and gentle procedure, the possibility to scale-up relatively easily, and compatibility with process automation. There is also a general consensus that MP offer a number of benefits in the context of immunoassay development compared to the utilization of two-dimensional structures (e.g. microtitter plates or chromatography membranes; Kuramitz 2009, Borlido et al. 2013, Rocha-Santos 2014).

The first one is a huge effective area. As a result, the incubation of a sample with MP in continuous suspension (i.e. under rotation or mixing) correlates with faster assay kinetics because the analytical target does not have to migrate very far and the probability of a binding event increases. Thanks to this, immunocapture can be performed in smaller sample volumes, in shorter assay times, and with lower LOD (Centi et al. 2005, 2007). Second is the possibility to finely tune MP surface to display different reactive groups or biomolecules and facilitate subsequent incorporation of bioreceptors. Third is the high chemical and physical stability, low toxicity, and high biocompatibility, making possible the immunocapture of a variety of targets, including live bacteria and eukaryotic cells. Fourth is the possibility to magnetically concentrate the MP after the immunocapture to carry out washes, serial incubations in different solutions, target preconcentration before detection or characterization studies, and/or MP confinement onto the transducer surface to supply signal amplification. Furthermore, MP can be easily released from a transducer’s surface after detection, facilitating sensor regeneration and reutilization compared to classical biosensing.

A good example is a work reported by Laczka et al. in which an immunoassay performed on MP detected Escherichia coli in a concentration range from 10⁴ to 10⁸ cells ml^-1 with an LOD of 3×10³ cells ml^-1, which was two orders of magnitude lower than the LOD obtained for the same Ab set used in a classical ELISA on microtitter plates (Laczka et al. 2011b). This assay was then formatted into a one-step electrochemical magnetoimmunoassay that provided detection of E. coli in a concentration range between 10² and 10⁸ cells ml^-1, with an LOD of 55 cells ml^-1, in about 1 h. Furthermore, carrying the immunoassay onto the MP allowed repeated reutilization of the electrodes and successful target detection in real sample matrices (i.e. diluted milk).

MP fabrication, handling, and immunofunctionalization

The basis behind MP utilization is simple. Because most biological samples display nonmagnetic behavior, a defined molecular target can be bound by MP, previously modified with a bioreceptor specific for that target, and be then separated from the other sample components by applying an external magnetic field. The whole procedure is relatively rapid and highly selective and occurs under very gentle conditions (low shearing forces) so that the analyte is not severely damaged for its subsequent study or detection.

For this to work, MP usually display superparamagnetic behavior. This implies that MP are composed or contain a superparamagnetic material that, by definition, exhibits zero remanence (and thus no magnetization) in the absence of external magnetic fields. Hence, MP are not magnetic “per se”. Thanks to this, MP are easily dispersed in solution and have low tendency to form aggregates by magnetic attraction to each other. On the contrary, when an external magnetic field is applied to the MP, a magnetic dipole is induced and there will be a net alignment of magnetic moments. The resulting magnetized MP will then be attracted towards the magnet. In this way, the supernatant can be removed by pipetting or aspiration for MP washing and resuspension in a different solution. If the external magnetic field is next removed or switched off, superparamagnetic MP return to their native nonmagnetic state and can be dispersed again in the solution.

Many methods have been described to produce MP and MNP (Xu and Wang 2012, Borlido et al. 2013). However, as a general rule, magnetic cores are first produced, mostly made of iron oxides such as magnetite (Fe₃O₄) and maghemite (γ-Fe₂O₃). These cores are subsequently coated or encapsulated (e.g. with a polymeric, protein, or silica coating) to provide stability, reduce metal leakage, facilitate bioengineering, and prevent the nonspecific binding of undesired components. Depending on the method used, the morphology of the MP will vary from encapsulated single magnetic cores, to groups of magnetic nucleus enclosed in a protein or polymeric matrix, or polymeric particles modified with multiplex magnetic nanocores on surface. The most notable exception has been for long Dynabeads, commercialized by Invitrogen (nowadays part of Thermo Fisher Scientific), which are among the most widely used commercial products. Contrary to other types of MP and MNP, Dynabeads are made of porous polymeric spheres evenly embedded with iron oxide. The resulting MP are 1 or 2.8 μm in diameter and are additionally covered with a polymeric layer to prevent leaking of the magnetic component and provide reactive groups for subsequent surface bioengineering. Dynabeads were the first MP released to the market, which were truly homogeneous in size and properties (Figure 4).

Figure 4:

Scanning electron microscopy (SEM) images of MP of different types and from different providers.

(A and B) Dynabeads, 2.8 and 1 μm, respectively, from Invitrogen (Thermo Fisher Scientific). (C and D) Multicore MP, 1 μm and 200 nm in diameter, respectively, produced by Chemicell.

The magnetic behavior of a specific type of MP depends on many factors, including composition, crystallinity of the structures, magnetic moment, and also the shape and size. Therefore, there is no general rule that can help predict the magnetic properties of MP. However, it is clear enough that, to capture MP from solution, the magnetic force exerted on them must be strong enough to overcome counteracting forces, such as diffusion, inertial, viscous, or gravitational forces. For most applications, using rare-earth magnets, which are permanent magnets significantly stronger than the classical ferrite ones, is sufficient. The strongest and more widely used ones are the neodymium magnets, which are made of an alloy of neodymium, iron, and boron (Nd₂Fe₁₄B), and provide magnetic fields of about 1.4 Teslas (compared to 0.5 Teslas generated by ferrite magnets). However, neodymium magnets only guarantee efficient recovery of micrometric MP and at small bench-top scales (up to 50 ml; Borlido et al. 2013). In the case of MNP, high gradient magnetic separators (HGMS) have to be used instead. These systems typically consist of two components, a canister filled with a magnetizable ferrous matrix (e.g. steel wool) and an electromagnet or a strong permanent magnet that provides a large external magnetic field. The resulting magnetic gradients are large enough to capture even weakly magnetic MNP in a flow stream (Borlido et al. 2013). Nevertheless, HGMS are only applicable to large process volumes, and there is still a lack of devices useful for the study of small samples or the confinement of small volumes of MNP on a transducer’s surface. This, together with the fact that MNP are more unstable and have a higher tendency to aggregate than MP, explains why, in spite of the high number of works reporting on the production and use of different types of MNP (Xu and Wang 2012), they have been used less often in the development of magnetoimmunoassays. In fact, most of the works published using SPE make also use of commercially available MP.

Ab incorporation onto MP surface can be accomplished by different means (Xu and Wang 2012). Because a significant number of MP with different functionalities on surface are commercially available, the protocol of choice often depends on biocomponent availability. For example, streptavidin, neutravidin, or avidin-coated MP can be used to bind biotininylated Ab through avidin-biotin affinity binding, which is characterized by extremely high affinity and specificity and is resistant to a wide range of temperature and pH. If the Ab preparation does not contain additional proteins or potentially interfering components (such as BSA or polymers), Ab can also be chemically conjugated to MP displaying reactive groups on surface. For this, several relatively simple chemical reactions are well known, which make use of commercially available cross-binding reagents. For instance, amidation reaction between amino/carboxyl functionalized MP and carboxyl/amino exposed on Ab can be carried out using reagents such as EDC/NHS. Chemical cross-binding between NH₂ groups on both beads and Ab can be done using glutaraldehyde. NH₂ groups in the Ab can be reacted with tosyl groups in tosyl-modified MP in a long but extremely simple one-step reaction. When these strategies are not possible, MP modified with protein A, protein G, or secondary Ab (e.g. anti-mouse Ab, anti-rabbit Ab) can be used to couple the Ab to the MP. This last alternative allows a certain degree of Ab orientation on surface, which might improve its performance and require lower amounts of Ab than chemical conjugation, and can be implemented even if the Ab preparation contains stabilizing components such as BSA or polymers. Once modified, the Ab-MP can be stored for months at 4°C without significant loss in performance (usually in saline solution supplemented with 0.1% BSA, 0.01% Tween-20X, and/or a preservative such as sodium azide).

MP in the development of electrochemical magnetoimmunoassays using SPE

For the electrochemical detection of magnetoimmunoassays, after immunocapture and washing, MP are confined onto the electrode surface using a magnet, which has to be placed immediately below the working electrode. The use of Ab-modified MP avoids in this way having to modify with Ab the surface of each single electrode, as it happens in classical biosensing, which is significantly longer and more tedious. Also, because MP can be easily released from the electrode by removing the magnet, electrode regeneration and reutilization is also easier and faster. Employing MP also provides a large active surface, both for Ab immobilization and for target immunocapture, which contributes to enhance target binding and generates larger signals than classical biosensing (Kuramitz 2009). This is particularly efficient when detecting analytes in complex sample matrices, in which target biomolecules may exhibit poor mass transport towards the biosensor surface, whereas nontarget components can contribute to its physical blockage by nonspecific adsorption (Xu and Wang 2012). In this context, MP play a role in molecular target preconcentration and separation from other nontarget sample components. In this respect, immunomagnetic concentration is a procedure that is usually simpler, milder, faster, and easier to perform, integrate, and automate than other established methods such as centrifugation, filtration, or chromatography.

In the specific case of electrochemical biosensing, using MP for target immunocapture guarantees that the incubations (i.e. with a potentially complex sample and other reagents) and the washing steps are carried out far away from the electrode where the electrochemical detection will take place. Accordingly, unmodified electrodes can be used for detection (in opposition to bioengineered devices in classical biosensing) and the surface of the working electrode is easily accessible by the enzymatic product, which diffuses onto the bare electrode surface without having to cross a protein biolayer (Baldrich et al. 2009). Besides, separating detection and bioassay on two surfaces facilitates immunoassay optimization and also the standardization of the electrochemical detection protocol. Hence, different magnetoimmunoassays can be processed following similar electrochemical detection procedures, which may help in subsequent integration and automation (Moral-Vico et al. 2015).

It is noteworthy that, whereas the examples of classical biosensors report LOD in the range between few picograms/milliliter and tens of nanograms/milliliter (Tudorache and Bala 2007, Mistry et al. 2014), many electrochemical magnetoimmunosensors described for analytes such as IgG, carcinoembryogenic antigen, α-fetoprotein, PSA, hepatitis B surface antigen, HIV antigen p24, or various hormones, among others, displayed LOD in the low picogram/milliliter range or lower (Xu and Wang 2012). For instance, Centi et al. compared the efficiency of two electrochemical immunosensing strategies for the detection of polychlorinated biphenyl (PCB) in food and environmental samples (Centi et al. 2007). Whereas the first one consisted of classical immunosensing onto the surface of a SPE modified with Ab, in the second case the immunoassay was performed onto MP and these were transferred to a bare SPE only for electrochemical detection. Both strategies were based on a direct competitive assay using alkaline phosphatase (AP) as the enzyme label, α-naphthyl phosphate as the enzymatic substrate, and DPV for detection. The sensitivity obtained by the magnetoimmunosensor was two orders of magnitude higher than that obtained with the classical immunosensor (LOD of 0.4 and 40 ng ml^-1, respectively), although it was also shorter (45 and 120 min, respectively). The authors concluded that the use of MP improved the yield of the affinity reaction, because the surface area of the solid phase was larger and the assay kinetics was faster. In a different approach, Esteban-Fernández de Ávila et al. developed an amperometric magnetoimmunosensor for human C-reactive protein (hCRP) and compared its performance to that of several commercial ELISA kits (Esteban-Fernández de Ávila et al. 2013). Whereas the biosensor had an LOD of 0.021 ng ml^-1, the spectrophotometric ELISA kits displayed LOD ranging between 0.002 ng ml^-1 (Abcam) and 100 ng ml^-1 (Abnova) for CRP. Although some commercial ELISA kits performed better than the amperometric magnetoimmunosensor, the sensor generated significantly lower LOD than many ELISA and produced the results faster than any of them. In the work reported by Torrente-Rodriguez et al., an electrochemical magnetoimmunosensor produced for the detection of interleukin-8 (IL-8) in saliva showed comparable performance than a commercial ELISA [LOD 26.4 and 25.0 pg ml^-1, linear range 87.9–5000 and 62.5–2000 pg ml^-1, and relative standard deviation (RSD) below 10%; Torrente-Rodriguez et al. 2016]. However, the sensor required 5 h to produce results, whereas the ELISA took 17 h.

Choosing the SPE

Currently, conductive inks of different composition and SPE made of different materials are commercially available (see, for example, Table 1). That is why the choice of the SPE material has to be made according to the electroactive target, label, or enzymatic substrate that will be detected. Nevertheless, many authors have observed that inks made of carbon tend to generate lower background currents than inks made of other materials, such as gold, and display also better electrocatalysis of a range of electroactive molecules and enzymatic substrates (Tudorache and Bala 2007). This implies that C-SPE often produce higher signal-to-noise ratios and lower LOD. In addition, C-SPE tend to be cheaper, which is an important consideration when buying high numbers of otherwise disposable devices. In agreement to this, the vast majority of the examples reported in the literature employed C-SPE (Table 3).

On the other hand, studies by Pernia et al. have shown that, when the distance between the working and counter electrodes is reduced, the sensitivity of the C-SPE increases proportionally (Pernia et al. 2009). We have also observed that, for a specific geometry and material, electrodes of smaller size provide lower signals but also lower background currents and LOD (Laczka et al. 2008b). Hence, the geometry of the SPE can play an important role in biosensor development, with devices with electrodes smaller and closer to each other performing better. It is also well known that parameters such as the exact ink formulation (e.g. type, size, or loading of graphite particles) and the printing and curing conditions, which are usually regarded by the manufacturer as proprietary information and are not provided to the user, can strongly affect the electron transfer reactivity and the overall analytical performance of the resulting SPE (Fanjul-Bolado et al. 2008, García-González et al. 2008). Hence, the selection of the best SPE for a certain application should optimally include the experimental testing of several SPE candidates, including, when possible, devices from different providers.

It is noteworthy that, although most providers claim that their SPE can be used in a straightforward manner when they are received, many authors have observed that pretreatment improves their performance to a certain extent, ameliorating electrochemical reversibility. For example, Wang et al. observed that C-SPE treatment by preanodization resulted in decreased peak-to-peak separation, increase in the registered signals, and lowering of the oxidation potential of a number of electroactive species, although also increase in the background current (Wang et al. 1996). Improved performance of C-SPE has been reported by other authors after anodization at an applied potential of 2.0 V vs. Ag/AgCl for a suitable time either in 0.1 m PBS, H₂SO₄, or NaOH and also after oxygen plasma treatment (Thiyagarajan et al. 2014). On the contrary, gold SPE may have to be activated, for instance, by repeated CV in H₂SO₄ until stable and reproducible voltammograms are obtained. In this respect, it is believed that electrode pretreatment helps to clean carbon and gold electrodes by removing organic binders and other insulating contaminants from the surface, by restructuring the resin matrix, by increasing surface roughening (thus effective area), and also by introducing defect/edge plane-like sites and surface-bound electroactive functionalities (such as quinone or carboxyl groups; Wang et al. 1996, Thiyagarajan et al. 2014). This is of special interest if the active groups that will be incorporated to the surface along the activation process are to be exploited for the subsequent conjugation of bioreceptors.

Selection of the MP

In general, spherical monodispersed MP will provide better hydrodynamic properties, more predictable magnetic behavior, and more homogeneous surface bioengineering and target binding than irregular MP showing a wide size distribution. On the contrary, multicore MP may be damaged easier during the incubation and washing steps due to mechanical stress. Independent of this, in applications requiring a high binding capacity, smaller particles will provide larger active surface areas. Furthermore, several reports describe that MNP provide faster and more efficient immunocapture rates than microscopic MP (Rocha-Santos 2014). However, MNP are more sensitive to irreversible agglomeration, especially in highly saline solutions such as many real sample matrices. In addition, because the magnetic force exerted on the MP decreases linearly with the product between its volume and magnetization, we have observed that MNP of size below 200 nm are not always efficiently attracted by normal magnets and have to be manipulated using special devices. Such an equipment is not available in every laboratory and is less compatible with magnetoimmunosensor development and high-throughput analysis. In contrast, whereas the magnetic properties of MNP can be improved by increasing their magnetic content, heavier beads sediment easier and faster, requiring more efficient mixing during the incubation steps.

Surface characteristic is another factor to consider, as it will condition the procedure for MP bioengineering and the types of target analytes that can be detected. In principle, it is accepted that highly hydrophilic MP surface coatings generate lower levels of biomolecule nonspecific adsorption, which is essential for the study of complex sample matrices. On the contrary, many providers sell MP with a variety of reactive groups displayed on surface, which facilitates the incorporation of Ab via different strategies.

MP immunofunctionalization

Ab incorporation onto the surface of MP can be accomplished by different means and the protocol and reagent of choice often depends on the Ab available (Xu and Wang 2012). For example, if biotinylated Ab can be obtained, using commercial (strept)avidin-coated MP will guarantee fast and efficient immunofunctionalization (i.e. usually a 15–30 min incubation with the biotinylated Ab followed by incubation with free biotin to block unreacted biotin-binding sites and prevent subsequent reagent cross-binding). If this is not the case but a relatively pure Ab preparation can be acquired at relatively high concentration (i.e. without additional proteins or potentially interfering components, such as BSA or polymers), Ab can be conjugated to the MP using chemical cross-linkers. The simplest and most widely used protocols are the amidation reaction using amino/carboxyl functionalized MP and EDC/NHS and glutaraldehyde cross-binding onto amino-MP. Some providers offer MP that display reactive surfaces, such as tosyl, N-hydroxy-succinimide, or epoxy groups that are able to react directly with -NH₂ or -SH groups on the Ab, which entails also an undemanding incorporation path. Ab chemical conjugation often requires that a chemical blocking step is included in the protocol (e.g. by incubating in the presence of amine-containing small molecules, such as Tris, lysine, glycine, or ethanolamine, or SH-containing molecules, such as cysteine). This makes sure that no reactive groups remain on surface, which could later react irreversibly with target and nontarget sample components.

When these strategies are not possible, but the Ab has been affinity purified against the target analyte (i.e. avoiding the direct use of antiserum and Ab purified by just protein A/G affinity column), MP modified with protein A/G or secondary Ab (e.g. anti-mouse for Ab produced in mouse or anti-rabbit for Ab produced in rabbit) can be used to couple the Ab. This last alternative allows a certain degree of Ab orientation on surface, which might improve performance, require lower amounts of Ab than chemical conjugation, and can be implemented even if the Ab preparation contains stabilizing components such as BSA or polymers. Nevertheless, and although protein A/G MP have been repeatedly used for the optimization of competitive immunoassay formats (Moreno-Guzman et al. 2010, Jodra et al. 2015a, b), these types of proteins will also cross-react with any IgG molecules present. This might be especially problematic when performing sandwich assay formats or when carrying out detection in mammalian serum. Ojeda et al. compared the performance of an electrochemical magnetoimmunoassay for the detection of ceruloplasmin, a copper-transporting α₂-glycoprotein present in serum, performed alternatively using streptavidin or protein A modified MP (Ojeda et al. 2013b). Both reagents allowed the optimization of competitive assay formats of comparable duration and performance. However, the authors found out that, whereas the assay on protein A MP displayed a linear response between 0.1 and 1000 μg ml^-1 and an LOD of 0.040 μg ml^-1, the assay produced using streptavidin MP had a linear range expanding between 0.025 and 20 μg ml^-1 and an LOD of 0.018 μg ml^-1. The RSD was also slightly better for streptavidin MP (3.2% compared to 5.2% for protein A MP). Although streptavidin MP needed slightly higher concentrations of Ab, they also suffered lower levels of interference by nontarget components in the serum samples than protein A MP.

Many MP producers provide modification protocols with their beads. However, the production of series of MP modified with increasing concentrations of Ab may help to determine the minimal amount of Ab required for efficient MP modification without unnecessary reagent waste. Following Ab incorporation, the surface of the MP has to be physically blocked to prevent subsequent nonspecific adsorption of nontarget components potentially present in the samples. The most widely used reagents for MP physical blocking are proteins (such as BSA, nonfat dry milk, casein, and gelatine) and polymers (such as dextran and ethylene glycol derivatives). However, and according to our own experience, a single blocking molecule might not display optimal performance for all the applications and should be carefully optimized. For example, when Ab-MP are produced for the study of complex real sample matrices, a short blocking with such a matrix (e.g. diluted blood serum, urine, or saliva) might help prevent subsequent nonspecific binding of nontarget components better than blocking with a single protein type, such as BSA. Blocking agents might play an additional role, as they act as a stabilizer for the Ab by minimizing the effects of denaturation caused by phase transitions associated with solid-phase binding. Independent of this, certain well-known blocking molecules should be avoided under some circumstances. For instance, BSA should not be used when studying samples potentially containing anti-phosphotyrosine Ab that could cross-react with the phosphotyrosine present in fraction V BSA preparations. Interference by BSA has been observed when the immunomodified MP had to be used to detect autoantibodies in saliva samples, in which salivary agglutinins produced important levels on nonspecific adsorption and cross-binding between the assay reagents and the sample components (Adornetto et al. 2015). Cross-reaction can be also observed when detecting small analytes if the Ab has been produced against a BSA-hapten conjugate (small molecules are typically linked to bigger protein carriers, such as BSA, to elicit an immune response as individual molecules). Another example is nonfat dry milk, a relatively inexpensive blocking agent, which may contain biotin that makes it inappropriate for use with (strept)avidin-biotin signal amplification systems and that may also contain inhibitors of the activity of the enzyme AP.

Magnetoimmunoassay optimization

Choosing the assay format for a magnetobiosensor primarily depends on the type of target molecule. For instance, a single immunobinding event might be sufficient for molecules that display electroactive or enzymatic behavior on their own (Figure 5A). Two good examples are dopamine (DA), an electroactive neurotransmitter, and myeloperoxidase (MPO), a peroxidase enzyme produced by blood neutrophiles (Baldrich and Munoz 2011, Herrasti et al. 2014a,b). In a completely different approach, Herrasti et al. proposed a generic assay format for one-step detection using MP modified simultaneously with bioreceptors and ferrocene (Herrasti et al. 2014c). They showed that binding of the target analyte produced physical sheltering of the labels and a concomitant decrease in the signal registered. To better detect the changes caused, they implemented CNT wiring of the MP surface, which enhanced 10-fold the signals measured compared to the direct detection of the MP alone. As a proof of concept, they accomplished the detection of a detergent and a biotinylated Ab, which bound to the beads by random physisorption and streptavidin-biotin affinity binding, respectively.

Figure 5:

Examples of magnetoimmunoassay format.

(A) Direct electrochemical detection of analytes that display electroactive or enzymatic behavior. Signal registered will be proportional to analyte concentration. (B) Direct electrochemical detection of nonelectroactive analytes using MP modified with Ab and an electroactive label. Signals inversely proportional to analyte concentration will be registered. (C) Sandwich immunoassay coupled to either direct detection (i.e. using an anti-target-labeled Ab) or indirect detection (i.e. using an unmodified anti-target IgG plus an anti-IgG-labeled Ab). (D) Three different formats of competition immunoassay.

Otherwise, more complex assay formats, such as the sandwich and competition ones, have to be implemented. Sandwich assay formats, in which a first Ab is used for capture and a second one is used for detection, are mostly used for the identification of whole cells, proteins, and other molecules big enough to display multiple binding sites (epitopes) on their surface (Figure 5C). Detection is accomplished by either using a detection Ab that has been modified with a detectable label (direct detection) or adding to the system a third Ab, which is a secondary-labeled Ab against the detection Ab (indirect detection). In both cases, target binding translates into signal generation and the signal will be proportional to the amount of target present in the sample. For the sandwich to work, an appropriate pair of Ab has to be selected, so that capture and detection Ab bind to different target epitopes and do not compete with each other. Optimally, one of the Ab will provide highly sensitive detection and the other will grant specificity. It is with this purpose that many sandwich immunoassays combine a polyclonal Ab (PAb) and a monoclonal Ab (MAb). Accordingly, the sandwich is the assay format that, generally speaking, produces the best results in terms of lowest LOD and highest specificity. On the contrary, it entails multiple incubation and washing steps, which makes it quite difficult to automate, especially to produce POC analytical devices. Several authors have attempted the combination of the different assay steps in a single incubation, which reduces drastically the number and duration of both incubation and washings (Laczka et al. 2008b, Kaçar et al. 2015, Moral-Vico et al. 2015). Nevertheless, such “one-step sandwich” immunoassays are only applicable to the study of real samples when the two Ab are highly specific for the target analyte and display high sensitivity for it, so that no interference occurs by either cross-binding or nonspecific binding of any nontarget molecules present.

In the case of detection of small target molecules that do not exhibit electroactive behavior, such as most drugs, hormones, pesticides, toxins, and other chemicals, a competition immunoassay scheme is the only applicable (Figure 5D). Here, a single incubation of the sample is performed in the presence of a competing molecule modified with a label, so that the target and labeled competitor molecules compete with each other for the binding sites in the MP. Therefore, the maximal signal is registered in the absence of the target molecule, whereas target presence induces a decrease in signal that should be proportional to its concentration. Because they usually involve a single incubation step, competition immunoassays are significantly faster than their sandwich counterparts and they can be extremely sensitive also. However, the specificity of a single immunocapture event strongly depends on the specificity of the Ab used and is, in most cases, lower than that of a sandwich, which implies two consecutive immunobinding events. Furthermore, the performance of a competition immunoassay arises from the differential affinity of the Ab for target and competitor. Considering that the production of labeled competitors is not necessarily easy and that the availability of commercial candidates is limited, assay improvement relies on a careful optimization of the competition conditions and might be limited for a given Ab/competitor set.

For any format type, the immunoassay has to be next optimized step-by-step. One of the parameters that should be studied is the amount of MP used per sample. In general, higher signals and faster assays are produced for increasing amounts of MP until saturation is reached, which reflects the higher availability of Ab and better mixing with the sample. Using more MP also seems to improve assay reproducibility, which has been attributed to enhanced magnetic recovery of the MP after the incubation and washing steps. On the contrary, confinement of too many MP onto the electrode surface might contribute to its partial sheltering and affect negatively the electrochemical transduction. Besides, several authors have reported that better detection of low analyte concentrations and lower LOD are attained with intermediate MP loads, which suggests that using as many beads as possible is not necessarily the best solution (Moreno-Guzman et al. 2010, Herrasti et al. 2014a). For example, Moreno-Guzman et al. studied amounts of MP ranging between 15 and 300 μg (equivalent to 0.5–10 μl of MP suspension per sample) for the immunodetection of cortisol, a steroid hormone (Moreno-Guzman et al. 2010). They observed that the peaks registered increased with the amount of MP up to 60 μg to decrease afterwards, which they attributed to the higher electron transfer resistance caused by large MP loadings on the electrode surface. However, they selected 30 μg MP as the optimal, which generated the highest signals for the lowest concentrations of cortisol tested, wider signal range, and a higher calibration slope value. In a different approach, Herrasti et al. assayed 30 to 50 μg MP (equivalent to 6–10 μl of MP suspension) for the detection of MPO (Herrasti et al. 2014a). They registered the highest signals for the lowest MP load (30 μg). They attributed this result to the fact that MPO is a dimeric protein, so that the use of more beads could led to MP cross-binding and aggregation, affecting negatively the enzyme activity.

Other parameters that can have an important effect on assay performance and should be optimized for each immunoassay are sample volume, immunocapture time and conditions, number of washings and washing buffer, or procedure for real sample dilution, with a wide variety of magnetoimmunoassay conditions having been employed. For example, sample volumes ranging from 50 to 100 μl are the most widely reported, with the exception of detection of bacteria and environmental pollutants, which are often studied in 0.5–1.0 ml sample volumes. In this respect, although increasing sample volume increases also the total amount of target molecules, it may also augment the amount of reagents needed and the assay final cost. MP recovery from larger volumes might also require longer times in the magnetic rack for efficient concentration and recovery. Most authors incubate the MP at room temperature, which produces lower levels of nonspecific adsorption than incubation at higher temperatures. However, we found few works reporting successful MP incubation at 37°C (Eguilaz et al. 2010, Moreno-Guzman et al. 2011, Ojeda et al. 2013a,b) and at least one work in which the authors claim 35% increase in the signals registered after incubating at 37°C compared to incubations at 25°C (Vidal et al. 2012). Because MP sediment over time, incubation has to be carried out under agitation conditions to keep the beads suspended in the solution. The different teams working in the field seem to use indistinct rotation using a mixing wheel, tilt rotation, and vortex-like shaking (such as using a thermomixer). Finally, series of washing steps with a detergent-supplemented buffer (such as PBS with 0.01–0.1% Tween-20) are usually implemented, with extended washes of up to 5 min each having been recommended to counteract high levels of nonspecific adsorption, especially for the study of complex sample matrices.

MP magnetic confinement onto the working electrode

Following immunobinding and washing, MP have to be confined onto the surface of the working electrode for electrochemical detection. This is accomplished by placing below a permanent magnet, most often a neodymium one. Many magnetic holders have been described in the literature for MP confinement onto thin-film electrodes, which are frequently too small for magnet manual placing. For example, one of us contributed to the development of a microfluidic device, which was produced using a combination of polymeric materials [a polycarbonate (PC) base, a polymethylmethacrylate (PMMA) cover, and a polydimethylsiloxane (PDMS) gasket] and was applied to the electrochemical detection of a sandwich immunoassay for E. coli bacteria quantization (Figure 6A; Laczka et al. 2011b). This device contained a set of thin-film microelectrodes placed in a microfluidic channel, a microchamber upstream from the electrodes, and a permanent magnet below the microchamber for MP confinement. In this way, MP were retained in the microchamber, whereas an appropriate enzyme substrate was flowed. The enzyme product was then pumped along the channel and was chronoamperometrically detected at the gold microband electrodes located downstream. The MP could be subsequently released by sliding a metal piece between the magnets and the chip, which allowed performing consecutive experiments. This design avoided direct contact of the biocomponents with the electrode, which lowered the risk of electrode fouling and facilitated electrode reutilization. In a successive work, this device was additionally improved to substitute the silicon-based thin-film electrodes by a low-cost plastic microfluidic cartridge that displayed four independent microfluidic channels and four electrode sets for the multiplexed electrochemical detection of cardiac biomarkers (Figure 6B; Moral-Vico et al. 2015). In that work, detection was alternatively carried out by registering steady-state currents under continuous enzyme substrate flow or by measuring the peak currents produced in a stopped flow approach (i.e. by stopping the substrate flow, so that the reaction product built up around the MP over time and could then be flowed downstream to the electrodes). As it was shown, appropriate tuning of the detection and flow conditions provided signal amplification and extremely sensitive detection. Furthermore, the implementation of different detection strategies on the different channels allowed also the simultaneous detection of assays or parameters that would produce signals of different orders of magnitude when measured by a single detection strategy. In a completely different approach, we produced a customized modular detector device using 3D laser sintering (i.e. a 3D printer; Figure 6C; Barallat et al. 2013). This tool integrated silicon-based thin-film electrodes, electrical connectors, and a novel magnetic switch, whose functioning was in this case founded on the vertical displacement of a permanent magnet. The design permitted that several devices were assembled to each other for multiplexed detection. As before, magnetic switching made possible the confinement of MP over the working electrode for electrochemical detection followed by MP release for electrode washing and reutilization.

Figure 6:

Examples of magnetic devices described previously for MP confinement onto a working electrode (additional information and bibliographic references are provided in the main text).

(A) Polymeric microfluidic device containing a set of thin-film microelectrodes placed in a microfluidic channel, a microchamber upstream from the electrodes, and a permanent magnet below the microchamber for MP confinement. MP are subsequently released by sliding a metal piece between the magnets and the chip. In this way, the physical separation of MP confinement and electrochemical detection avoids direct contact between biocomponents and electrodes, preventing fouling and facilitating reutilization of the electrodes. (B) A related device, this time incorporating a low-cost plastic microfluidic cartridge that displays four independent microfluidic channels and electrode sets for multiplexed detection. (C) Modular detector device produced by 3D laser sintering. It incorporates a magnetic switch based on the vertical displacement of a permanent magnet. (D) SPE displaying a magnet that has been pasted with cello tape below the working electrode. (E) Images of the magnetic devices provided by Dropsens for their 8-SPE and 96-SPE arrayed devices (images kindly provided by Dropsens). (F) Image of the PalmSens magnetic device specially designed for their 8-SPE cartridge (image kindly provided by PalmSens). (G) Customized magnetic device produced in PMMA by standard milling and laser cutting. It consists of a base that can house up to eight SPE, a sliding component with eight magnets for directed and reversible MP confinement, and a cover with eight wells that allow electrode washing and MP/solution delivery with minimal risk of spilling and cross-contamination.

An important proportion of the works dealing with SPE do not state how the magnet is hold in place below the electrode. However, the big size and translucent substrate that characterizes many SPE makes it possible to stick a magnet under the working electrode using just a cello tape, an extremely low-cost strategy that we have extensively used in the past (Figure 6D). Alternatively, the magnet(s) can be embedded in a physical substrate made of plastic or silicone, which is then slid or placed under the SPE. This facilitates very much the operation and handling of multiplex SPE. A few of such magnetic devices are commercially available (Figure 6E and F). In a slightly different and more sophisticated approach, we recently produced a device with switchable magnets, which was simple, cheap to produce, and very easy to use (Figure 6G; Baldrich and del Campo 2015, García-Robaina et al. 2016). This device was fabricated in PMMA combining standard milling and laser cutting. It consisted of a base that could house up to eight SPE, a sliding component with eight magnets for directed and reversible MP confinement, and a cover with eight wells that allowed electrode washing and MP/solution delivery with minimal risk of spilling and cross-contamination.

Immunoassay electrochemical detection

Concerning detection, the majority of the electrochemical magnetoimmunoassays reported in the literature make use of enzyme labels and appropriate enzyme substrates. The enzyme most widely used is horseradish peroxidase (HRP), especially in the competition assay formats. This might be related to the fact that HRP is extracted from horseradish roots and can be obtained at relatively low price. In addition, HRP is a relatively small (about 45 kDa) and very stable enzyme that survives chemical modification quite well. Numerous substrates have been described for the spectrophotometric, fluorescent, and electrochemical detection of HRP. Electrochemical detection is carried out in most of the cases using a mixture of hydroquinone (HQ) and hydrogen peroxide (H₂O₂) as the substrate solution, where the enzyme catalyzes the reduction of H₂O₂ coupled to the oxidation of HQ into benzoquinone (BQ). BQ is then reduced back to HQ at the electrode surface and the reduction current produced is correlated to the amount of enzyme present. Because HQ and BQ are very toxic for both health and the environment, many alternatives have been assayed. The two best known alternatives are 3,3′,5,5′-tetramethylbenzidine (TMB) and 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), which can be also obtained as ready-to-use solutions optimized for HRP colorimetric detection. We have managed to use some of these ready-to-use products for electrochemistry with acceptable results and higher reproducibility than when using a fresh homemade preparation for each single detection experiment (Herrasti et al. 2014a, Moral-Vico et al. 2015). Although TMB is known to produce electrode fouling and partial passivation over the measurement, it has been successfully used for magnetoimmunoassay detection at disposable SPE with results comparable to those of HQ. ABTS, on the contrary, produces signals considerably lower than HQ and TMB for HRP detection.

The other enzyme that has been extensively employed for the electrochemical detection of magnetoimmunoassays is AP, in most of the cases, using α-naphthyl phosphate as the substrate. The reaction product is in this case oxidized at the electrode surface. Although the literature lacks detailed comparisons, the results reported by the different teams for the detection in saline buffer suggest that not a single enzyme label or enzyme-substrate combination performs consistently better than the others and that immunoassay sensitivity is probably more related to appropriate Ab selection and assay optimization (Table 3). For instance, whereas works based on the use of AP and p-nitrophenol describe LOD ranging from 5 pg ml^-1 to 3.74 ng ml^-1, those based on HRP and HQ/H₂O₂ report values between 0.5 pg ml^-1 and 11 ng ml^-1.

Detection of biomarkers in samples of biological origin

Monitoring the levels of certain small hormones in urine and serum is relevant for, among others, the control of doping in sport competitions. Chromatographic methods, such as high-performance liquid chromatography (HPLC), gas chromatography-mass spectrometry (GC-MS), and liquid chromatography-mass spectrometry (LC-MS), are the most widely used for this purpose and they are known to provide quantitative and confirmatory results with high sensitivity and selectivity. However, all of them rely on the employment of sophisticated and expensive instrumentation and suffer from considerable time delay between sampling and result generation. Accordingly, the availability of electrochemical immunosensors, easy to use, potentially miniaturized, and yet able to afford high sensitivity and reproducibility, is envisaged as a promising alternative for doping monitoring. Immunosensing of such small-sized molecules has to rely on competitive assay formats. In this context, researchers from Pingarron’s group, one of the pioneers in the field, described an electrochemical immunosensor for testosterone. It was based on a direct competitive immunoassay format using MP modified with protein A and MAb, a testosterone-HRP conjugate, and amperometric detection using HQ/H₂O₂ and C-SPE (Eguilaz et al. 2010). This is one of the few works in which magnetoimmunocapture has been performed at 37°C instead of at room temperature. The sensor displayed a linear range expanding between 5 pg ml^-1 and 50 ng ml^-1, an LOD of 1.7 pg ml^-1, and RSD values below 10%. The sensor was also used to detect testosterone concentrations between 1 and 10 ng ml^-1 spiked in serum, in which it showed 10.4% and 0.8% of cross-binding against 19-nortestosterone and estradiol, respectively. The authors demonstrated that the immunomodified MP were stable for at least 25 days stored at 4°C. In a later work, the same team described an immunosensor for the determination of cortisol (Moreno-Guzman et al. 2010). In this case, anti-cortisol MAb were immobilized onto protein A MP and a competitive immunoassay was performed at 37°C using AP-labeled cortisol as the competitor. The electrochemical detection was carried out by DPV using 1-naphthyl phosphate as the AP substrate. The sensor displayed a linear range between 5.0×10^-3 and 150 ng ml^-1, with an LOD of 3.5 pg ml^-1, and could produce an accurate estimation of the concentration of cortisol present in two certified human sera after appropriate dilution.

Not all hormones are small molecules and a number of them are peptides or proteins of a certain size that can be detected using classical sandwich immunoassay formats. Some of them are regularly used as biomarkers of particular disease conditions. For instance, prolacting is a polypeptide hormone (m.w. 23,000 Da) that, among its various functions in the body, stimulates lactation and displays regulatory roles in the growth and differentiation of the mammary gland. Abnormal concentrations of prolactin in women serum may be indicative of menstrual disturbances, hypogonadism, or infertility, among other diseases. A disposable electrochemical magnetoimmunosensor for prolactin was developed based on a sandwich immunoassay performed at streptavidin-coated MP using MAb (Moreno-Guzman et al. 2011). The electrochemical detection was performed by DPV using C-SPE, AP as the enzyme label, and 1-naphtyl phosphate as the enzymatic substrate. According to the authors, the sensor displayed linear response in the 10 to 2000 ng ml^-1 concentration range, with nearly indistinguishable slopes in either standard solutions or spiked serum, LOD of 3.74 ng ml^-1, and RSD below 9% for measurements performed on different days. Moreover, once prepared, the anti-prolactin MP were stable for at least 34 days if they were stored at 4°C [coefficient of variation (%CV), 6%].

Another example is leptin, a 16-kDa peptide hormone that plays an important role in regulating food intake and body composition in mammals. It has been observed that the concentration of leptin is higher in obese individuals, and it has been suggested that leptin present in breast milk might protect against overweight in adulthood, which could partially explain the increased risk of obesity of formula-fed infants with respect to breastfed infants. Ojeda et al. developed an electrochemical magnetoimmunosensor for leptin quantification using modified streptavidin MP, a biotinylated anti-leptin capture Ab, a detection anti-leptin MAb, and its indirect detection with an AP-labeled anti-mouse (Ojeda et al. 2013a). The electrochemical detection was accomplished by DPV at C-SPE using 1-naphthyl phosphate as the enzyme substrate. The assay displayed linear response between 5 and 100 pg ml^-1, an LOD of 0.5 pg ml^-1, and intraday and interday RSD values below 8%. Besides, the immunomodified MP were stable at 8°C for 30 days. The usefulness of the immunosensor was demonstrated by analyzing human serum, infant powdered milk, and breast milk from a nursing mother, all of them spiked with known concentrations of leptin and diluted 1:1000 before analysis, with recoveries ranging between 96% and 102%.

Apart from hormones, many peptides and proteins are used as biomarkers of different pathological states. Numerous analytical methods have been applied to the detection, study, and quantization of peptide and protein biomarkers, such as capillary electrophoresis, radioimmunoassay, Western blotting, and ELISA, to cite just a few. Among them, ELISA, carried out either manually or using robotized platforms at centralized laboratories, is one of the best known gold standards. In this specific field, immunosensors are expected to offer faster detection, simpler assay formats easier to perform by a non-specially trained personnel, lower sample volume requirement, and lower LOD. For example, Esteban-Fernández de Ávila et al. described an amperometric magnetoimmunosensor for the detection of the cardiac biomarker amino-terminal pro-B-type natriuretic peptide (NP-proBNP) in human serum (Esteban-Fernández de Ávila et al. 2013). In this case, the Ag was chemically conjugated onto MP displaying -COOH groups, and the assay consisted of a competition between anti-NP-proBNP Ab modified with HRP and native NP-proBNP in the samples. The amperometric detection was carried out at gold SPE using TMB/H₂O₂. According to the results, the sensor’s dynamic range expanded from 0.63 to 77.30 ng ml^-1, with an LOD of 0.21 ng ml^-1 (calculated as the analyte concentration for which the maximum amperometric signal was reduced by 10%). The sensor was then calibrated by detecting NP-proBNP concentrations ranging from 0.01 to 5000 ng ml^-1 in 10 times diluted spiked human serum samples. Under these conditions, the dynamic range was between 0.12 and 0.49 ng ml^-1, with an LOD of 0.02 ng ml^-1, in a 45-min assay format. In a different approach, Torrente-Rodriguez et al. produced an electrochemical bioplatform for the simultaneous determination of IL-8 mRNA and IL-8 protein, two oral cancer biomarkers, in raw saliva (Torrente-Rodriguez et al. 2016). For detection, they employed dual C-SPE composed of two elliptic carbon working electrodes (6.3 mm² each), a carbon counter electrode, and Ag pseudo-reference electrode. Protein binding was performed using MP 2.8 μm in diameter based on a sandwich assay format coupled to the amperometric detection of the HRP enzyme label using HQ/H₂O₂. When the authors compared the performance of their immunosensor to that of the commercial ELISA that was being used as the reference method, both produced comparable LOD (i.e. 26.4 and 25.0 pg ml^-1), linear ranges (i.e. 87.9–5000 and 62.5–2000 pg ml^-1 for sensor and ELISA, respectively), and RSD (<10%). However, whereas the sensor entailed 5 h, the ELISA took 17 h.

Celiac disease is a gluten-induced autoimmune enteropathy that develops in genetically susceptible individuals. It is one of the most common immune-mediated diseases in Europe and North America, where it is estimated to affect approximately 1% to 1.2% of the population. Celiac disease can be diagnosed in blood serum samples by detecting three types of specific biomarkers, including anti-transglutaminase IgA autoantibodies (anti-TG2), anti-endomysial autoantibodies, and Ab against deaminated forms of gliadin (one of the proteins composing cereal gluten). Kergaravat et al. reported an electrochemical magnetoimmunosensor for the detection of anti-TG2, which is crucial for the diagnosis and monitoring of celiac disease (Kergaravat et al. 2013). In this case, transglutaminase Ag was immobilized on tosyl-activated MP. Following incubation with serum samples, MP were incubated with an HRP-labeled anti-human Ab. The detection of the enzyme labels was performed by SWV using o-phenylenediamine and H₂O₂ at two different types of electrodes. The first one consisted of a graphite-epoxy composite cylindrical electrode with an embedded magnet, which was used with an external platinum wire as the auxiliary electrode and a silver/silver chloride (Ag/AgCl) as the reference electrode. The second one was a C-SPE. Although both devices performed well, the responses registered at the C-SPE were in average five times higher. This sensor was used to study a total of 29 sera from clinically confirmed cases of celiac disease and 19 negative control sera. The results correlated well with those of a commercial ELISA, allowed effective discrimination between the two types of samples, and indicated 100% sensitivity and 84% specificity. A similar approach has been recently exploited by Adornetto et al. for the noninvasive detection of celiac disease in saliva samples. In this occasion, the authors used AP as the label, 1-naphthyl phosphate as the enzyme substrate, DPV for measuring, and a strip of eight magnetized SPE as the electrochemical transducer (Adornetto et al. 2015). The immunoassay had to be extensively optimized to function in saliva, where high amounts of salivary agglutinins provided unacceptable levels of nonspecific binding and reagent cross-binding. With this objective, the BSA in the blocking and washing buffers was substituted by bovine gelatine, and the washing buffer was also supplemented with higher amounts of detergent and a chelating agent (5 mm EDTA and 0.3% Tween-20). The optimized immunosensor was finally applied to the study of 66 saliva samples that had been independently classified as positive or negative using a radioimmunoassay method. According to the authors, the sensor provided 95% clinical sensitivity and 96% clinical specificity.

Environmental monitoring

The examples of magnetoimmunoassays developed for environmental monitoring have been mostly focused on the detection of small contaminant molecules. For example, an immunomagnetic electrochemical sensor for the determination of 2,4-dichlorophenoxyacetic acid in river water samples was developed by Dequaire in 1999 (Dequaire et al. 1999). The assay consisted of a competition using an AP conjugate. For detection, an array of four C-SPE was printed on a flexible polyester film (i.e. a film for overhead projection), the working electrodes were coated with Nafion, and a polystyrene cylinder was assembled to produce microwells onto each device. The authors declared that the sensor displayed a working range between 0.01 and 100 μg l^-1, an LOD of 0.01 μg l^-1, RSD of 9%, and negligible interference by other sample components. Some years later, Centi et al. optimized a competitive immunoassay format for the detection of PCB in marine sediment extracts (Centi et al. 2005). Electrochemical detection was carried out using AP and α-naphthyl phosphate in two different assay formats. On the one hand, a classical immunosensor was produced by immobilizing the Ab on the surface of a C-SPE. On the other hand, immunobinding was carried out on immunomodified MP that were confined onto the C-SPE just for detection. The second format displayed an LOD 100 times lower than the classical immunosensor (40 and 0.4 ng ml^-1, respectively) and produced results in shorter assay times (125 vs. 45 min). In view of the results, the authors concluded that the magnetoimmunosensor guaranteed detection of lower concentrations of PCB, with high reproducibility and nearly undetectable interference by nontarget sample components.

Examples of detection of whole bacterial cells have been also described. Rapid Legionella pneumophila determination was accomplished based on a disposable core/shell Fe₃O₄@poly(DA) MNP immunoplatform (Martín et al. 2015). The quasi-spherical MNP had a magnetic core of 12.5±0.9 nm and a DA-based polymeric thin layer of about 4.0±0.6 nm and were modified with a specific PAb. For the amperometric detection, the authors used HRP-Ab, HQ/H₂O₂, and C-SPE. The sensor showed linear response against the logarithmic value of L. pneumophila cell concentration over four orders of magnitude [between 1.0×10⁴ and 1.0×10⁸ colony-forming units (CFU) ml^-1] and an LOD of 1.0×10⁴ CFU ml^-1. Despite the good analytical performance of the developed methodology, the authors observed that the LOD obtained was not low enough to allow the detection of L. pneumophila in cooling towers, where the reference levels are less than 100 CFU l^-1. For this reason, the procedure was additionally improved by implementing a membrane-based preconcentration step. This made possible the detection in a concentration range from 10 to 10⁵ CFU ml^-1, with an LOD of only 10 CFU ml^-1. In addition, the sensor showed no significant interference by unrelated nontarget microorganisms, such as E. coli, Enterococcus faecalis, Pseudomonas aeruginosa, Acinetobacter baumannii, Aeromonas, and Salmonella, and took less than 3 h compared to the 10 days required for the standard culture methods.

Food safety monitoring

Mycotoxins are secondary metabolites produced by different filamentous fungi. Although some mycotoxins, such as penicillin, have been exploited with medical purposes, others may cause disease and death in humans and many other animals. It is noteworthy that mycotoxins are very resistant to decomposition and are not severely damaged during digestion or temperature treatment. For this reason, they may remain in meat and dairy products even after cooking and/or freezing. Hence, the entrance of infected crops in the food chain, either via human nourishing or livestock feed, can have an important impact both in health and economical terms. Alberto Escarpa’s team has worked extensively in the development of biosensors for the detection of different mycotoxins, some of them using MP and SPE. For example, Hervás et al. described an electrochemical magnetoimmunosensor for the determination of zearalenone in baby food (Hervás et al. 2009). It consisted of a direct competitive immunoassay and amperometric detection of the HRP labels at C-SPE using HQ/H₂O₂ as the enzyme substrate. Although the method displayed high reproducibility (RSD of 7.8%), the authors observed important levels of matrix interference when they analyzed complex samples such as cereal milkshakes. For instance, whereas an LOD of 0.011 μg l^-1 was obtained in saline buffer, the LOD was 0.12 μg l^-1 in cereal milkshakes. This was attributed by the authors to the presence of organic solvents, lipids, vitamins, or proteins in the sample or the extracts that could disrupt the interaction between Ag and Ab. This study was subsequently extended in a work in which the electrochemical detection was improved by performing DPV in lower substrate volumes, which allowed generating an LOD of 0.007 μg l^-1 in shorted detection times (5 min of incubation with HQ/H₂O₂ instead of 25 min; Hervas et al. 2010). The same team reported more recently on electrochemical magnetoimmunosensors based on a similar methodology for the determination of fumonisins in maize and beer (Jodra et al. 2015b) and ochratoxin A (OTA) in coffee (Jodra et al. 2015a). According to the authors, the use of MP and disposable C-SPE, coupled to the implementation of a simultaneous simplified calibration and analysis protocol, provided a fast and reliable method for mycotoxin detection in complex real matrices. Furthermore, dilution of the samples 1:5 or 1:10, depending on the sample type, seemed enough to prevent any interference by other nontarget components. On the contrary, the study of undiluted beer samples correlated with a 32% decrease of the amperometric current compared to the signals registered in buffer (Jodra et al. 2015b).

Vidal el al. reported also an electrochemical immunosensor for OTA determination, in this case in wine (Vidal et al. 2012). In this work, streptavidin-coated MP (1 μm diameter) were modified with a biotinylated MAb and a competition assay was performed using an OTA-HRP conjugate tracer. The electrochemical detection was carried out by DPV using homemade C-SPE and HQ/H₂O₂ as the substrate solution. The sensor displayed linear response for OTA concentrations ranging from 0.15 to 8 ng ml^-1 and an LOD of 0.12 and 0.11 ng ml^-1 in PBS and wine, respectively. These values were comparable to the LOD obtained for HPLC, which was used as the reference method (about 0.030 ng ml^-1), and below the concentration limits established by the European Commission for wine samples (<2 ng ml^-1). It is noteworthy that the sensor required smaller volumes of reagents and lower incubation and measurement times (total assay time of 40–50 min, excluding extraction but including 30 min for OTA immunocapture).

Advantages and drawbacks of signal amplifiers and signal amplification strategies

The majority of the electrochemical magnetoimmunosensors reported to date lean on the utilization of enzyme labels. In this context, the oldest and simplest signal amplification strategy has been for long the utilization of a biotinylated Ab (Ab-biotin) for detection, followed by incubation with a (strept)avidin-bound enzyme (Guesdon et al. 1979, Savage et al. 1992). The (strept)avidin-biotin technology relies on the extremely high and specific affinity of (strept)avidin for biotin (K_d≈10^-14–10^-16m). (Strept)avidins are also exceptionally stable proteins. In addition, using indirect (strept)avidin-biotin detection strategies, the experimentalist uses an Ab-biotin that has suffered a less severe modification than an enzyme-conjugated counterpart. Hence, lower amounts of biocomponents are usually needed, which also provide improved performance. Besides, each Ab-biotin can be bound by more than one (strept)avidin molecule, which may display on surface more than one enzyme unit. Accordingly, this indirect detection strategy typically results in signal amplification compared to direct detection using an Ab-enzyme. On the contrary, indirect detection implies extending the assay with additional incubation and washing steps compared to direct detection using an Ab-enzyme conjugate, without forgetting that signal amplification works for both specific and nonspecific signals. Nevertheless, some of the drawbacks classically related to these systems, such as the high levels of nonspecific adsorption typically reported for avidin, have been solved over the last years as new engineered (strept)avidin molecular derivatives have been produced and commercialized. For example, NeutrAvidin and ExtrAvidin are deglycosylated versions of avidin that display neutral isoelectric point (pI), and reportedly produce lower levels of nonspecific binding and undetectable binding of lectins compared to avidin (Laitinen et al. 2006, 2007). In addition, (strept)avidin variants that display lower affinity for biotin, such as CaptAvidin, are also available and have been tested for the production of regenerable and reusable biosensors (Campbell and Mutharasan 2005, Garcia-Aljaro et al. 2009, Munzer et al. 2014).

In recent years, there has been increasing interest in using nanomaterial tags as multilabel carriers, enzyme mimetics, and nonenzymatic labels (Ding et al. 2013, Hu et al. 2013, Huang and Zhu 2013, Wei and Wang 2013, Shen et al. 2014). Some examples involve noble metal nanoparticles, carbon nanomaterials, semiconductor nanocrystals (or QD), metal oxide nanostructures, and hybrid nanostructures, which in some instances have provided LOD 10 times lower than those registered using classical Ab-enzyme conjugates (Ding et al. 2013, Huang and Zhu 2013, Pei et al. 2013). Compared to classical Ab-enzyme conjugates, in which 1:1 molar ratios are usually approached, nanoparticles display high surface-to-volume ratios. Besides, the production and modification of nanomaterials can be tailored to control their final properties and characteristics. Consequently, highly efficient nanotags can be produced that exhibit on surface elevated numbers of Ab, increasing the possibility of Ab-Ag binding, as well as numerous label units able to provide signal amplification. In this respect, either enzymes such as HRP and AP or electroactive molecules such as thionine, ferrocene derivatives, and methylene blue have been employed. Alternatively, redox-active nanostructures have been used as enzyme mimetic catalysts. For example, many authors have defended that gold nanoparticles (AuNP) can catalyze the reduction of p-nitrophenol to p-aminophenol in the presence of NaBH₄ and can thus be used as an HRP mimic (Das et al. 2006, Lin and Doong 2011). Although, for this approach, the addition of an enzyme substrate is required, enzyme mimetics have been reported to be more stable than their enzymatic counterparts under a wide variety of conditions (e.g. temperature and pH). Finally, many nanomaterials, such as metal nanoparticles and QD, are also conductive and can be detected electrochemically.

In spite of the many advantages described, the utilization of nanocomponent nanotags faces also a number of drawbacks (Huang and Zhu 2013). For instance, most nanomaterials have a strong tendency to aggregate in saline solutions, which make difficult their direct utilization in biological sample matrices. On the contrary, many QD are made of potentially toxic heavy metal ions (e.g. Cd²⁺), which may hamper their practical application. Although different surface coating strategies have been reported to provide nanolabels with higher stability and biocompatibility and lower toxicity, such modifications often affect also tag optical and electrochemical behavior. The electrochemical detection of QD is also affected by pH, with maximal signals often registered at slightly acidic pH. Finally, many nanoparticles are electroinactive within the potential window available in aqueous electrolyte media or can only be oxidized/reduced in the potential region close to the solvent potential limit where the background process is dominant. This outcome is often associated with their intrinsic stability, especially when they have been strongly capped with stabilizers. This means that it is often difficult to detect nanoparticle labels directly using electrochemical methods. For this reason, although few examples have reported straightforward detection, the electrochemical detection of metal nanoparticles and QD is usually accomplished by stripping voltammetry after metal dissolution in acidic solution, which in fact entails the addition of a reagent and cannot be truly considered a reagent-less detection strategy (Ding et al. 2013). Another factor that should be taken into account is the effect of nanotag nonspecific binding. Because nanomaterials provide important levels of signal amplification, their nonspecific interaction with either MP, target molecules, or the SPE surface might produce significant signals and ultimately result in biosensor lack of reliability (Ding et al. 2013). Accordingly, additional research will be needed to optimize nonfouling surface treatments for both the nanoparticle labels and MP.

Electrochemical detection of magnetoimmunoassays based on (strept) avidin-biotin binding

Streptavidin-HRP was used for detection in the work reported by Eletxigerra et al., who developed an amperometric magnetoimmunoassay for the detection of tumor necrosis factor-α (TNF-α) in human serum (Eletxigerra et al. 2014). After optimization of the sandwich magnetoimmunoassay, LOD of 2.0 pg ml^-1 (36 fM) and 5.8 pg ml^-1 (105 fM) were obtained for standard solutions and spiked human serum samples, respectively. These values lay within the clinically relevant range, and elevated TNF-α biomarker levels could be detected. According to the authors, the improved analytical performance of their magnetoimmunoassay was partly attributable to the careful reoptimization carried out in the real sample matrix. On the contrary, the authors were aware that the main drawback of the assay presented was the need to perform two incubations of 1 h each. They suggested that reducing the number of steps of the sandwich immunoassay would reduce the assay time, although at the expense of slightly increased LOD. A similar concern was discussed in a work presented by Pingarron’s team, in which an ultrasensitive amperometric magnetoimmunosensor for hCRP quantification in serum was reported making use of streptavidin-HRP as a signal amplifier (Esteban-Fernández de Ávila et al. 2013). Here, the sandwich magnetoimmunoassay consisted originally of three incubation steps, which were subsequently combined by the authors in two and a single incubation step using a multireagent cocktail. The conclusions of this study showed that, as the number of steps diminished, the assay signal-to-blank ratio decreased also. Therefore, the authors decided to continue working with the long three-step protocol, which produced a very low LOD (0.021 ng ml^-1). As a comparison they studied several commercial ELISA kits, which displayed LOD ranging between 0.002 and 100 ng ml^-1. The authors concluded that, apart from the lower LOD achieved by the sensor, the ELISA required the pretreatment of the biological sample (normally centrifugation) and involved complicated, tedious, and time-consuming multistage processes.

Electrochemical detection of metal nanoparticle-based nanotags

The electrochemical detection of metal nanoparticles is usually accomplished by stripping voltammetry after metal dissolution in acidic solution. However, attempts to provide direct detection of AuNP, without prior chemical dissolution, have been also reported. For example, an AuNP-based electrochemical magnetoimmunosensor for the rapid detection of anti-hepatitis B virus Ab in human serum was described (de la Escosura-Muñiz et al. 2010). The electrochemical detection of the nanolabels was carried out by monitoring chronoamperometrically their catalytic properties towards the hydrogen evolution in an acidic medium. However, this work included only a limited amount of results that were not compared to any standard detection methodology, such as detection based on HRP. The same team exploited later a similar procedure for the highly sensitive and rapid determination of E. coli O157:H7 in minced beef and water using electrocatalytic AuNP tags (Hassan et al. 2015). Anti-E. coli O157 MBs were used as a capture platform and Ab-modified AuNP were employed in a sandwich assay format and were detected by chronoamperometry at C-SPE. According to the authors, the electrochemical assay displayed specificity and reproducibility comparable to those of a commercial lateral-flow kit but better detection range. Furthermore, the LOD obtained using the sensor in saline solution, minced beef, and tap water samples were 1000-, 200-, and 300-fold better, respectively. It is noteworthy that the results shown for the electrochemical determination of E. coli O157:H7 in PBS include quite a high background current that is not mentioned by the authors, which is compatible with the important levels of nanolabel nonspecific adsorption described by other authors.

One of the main drawbacks of the use of AuNP in electrochemical assays is that gold dissolution depends on the addition of harsh oxidizing agents, such as hydrobromic acid/bromine. To overcome the use of these agents, Szymanski et al. introduced a method in which silver nanoparticles (AgNP) and MP were used as a platform for immunosensing (Szymanski et al. 2011). In this work, the simplicity and analytical utility of AgNP used as immunolabels was illustrated by detecting myoglobin across a range of concentrations of physiological interest, which permitted distinguishing potential myocardial infarctions from normal background levels in serum. For the detection of AgNP, an oxidizing potential was applied to the C-SPE, which was enough to dissolve the silver without the need for an external oxidizing agent.

Although most of the examples reported made use of MP, a few authors have highlighted the benefits of using MNP instead as a promising material for enhancing the sensitivity of electrochemical immunoassays. In particular, hybrid immuno-MNP, consisting of two or more different nanoscale functionalities, have attracted much attention due to their novel combined properties and multiple potential applications. A good example are Fe₃O₄ (core)/SiO₂ (shell) MNP, which combine well-known surface functionalities for the easy and versatile immobilization of Ab, with efficient magnetic separation from unbound proteins. For instance, Gan et al. developed an ultrasensitive electrochemical immunosensor for HIV p24 based on core/shell Fe₃O₄@SiO₂ MNP and a nanogold colloid-labeled enzyme-Ab copolymer as the signal tag (Gan et al. 2013). They concluded that the developed disposable amperometric immunosensor could detect picogram/milliliter concentrations of HIV p24 protein, which was 1000 times below the LOD of a routine ELISA method. In another work, Fei et al. exploited the special properties of MNP labeled with AuNP (AuNP/SiO₂/Fe₃O₄; Fei et al. 2015). The authors developed a sandwich electrochemical immunoassay for Salmonella pullorum and Salmonella gallinarum based on Ab-coated AuNP/SiO₂/Fe₃O₄ and four-channel C-SPE modified with electrodeposited AuNP. One of the novelties of this work was the use of the multiplexed four-channel C-SPE, which showed higher reproducibility than four independent C-SPE. This was partly attributed to the fact that the four-channel C-SPE used the same reference and auxiliary electrodes, which minimized interference by variable external factors. Another contribution of this work was the dual implementation of AuNP. On the one hand, AuNP were used as bridging materials between the biomolecules and the Fe₃O₄/SiO₂-SH MNP, because the AuNP can easily immobilize the Ab while retaining a high bioactivity of the adsorbed biomolecules. On the other hand, the AuNP electrodeposited on the working electrodes increased the signals registered. As a result, the sensor displayed a wide linear range, low LOD, and high specificity.

Using QD for immunoassay electrochemical detection

Semiconductor QD have received considerable attention because of their unique properties, including high quantum yield, simultaneous excitation with multiple fluorescence colors, and electrochemical properties. Accordingly, QD have been extensively used as both fluorescent and electrochemical labels. One of the most promising characteristics of QD electrochemical detection is the possibility to carry out multiplex detection. For instance, in a pioneer work, Prof. Wang’s team reported on an electrochemical sandwich immunoassay for the simultaneous detection of four different target proteins. The assay employed four different types of semiconductor tags made of cadmium sulfide, zinc sulfide, copper sulfide, and lead sulfide, which had been modified with four different Ab (Liu et al. 2004). In this work, detection was accomplished by stripping voltammetry of the corresponding metals dissolved in acid.

QD have been also used in combination with MP in electrochemical assays. In a work by Wang et al., an electrochemical magnetoimmunoassay with QD labels was produced for the detection of phosphorylated acetylcholinesterase in plasma (Wang et al. 2008). According to the authors, the good performance of the sensor was due to a number of factors combined. First, to the selection of a pair of Ab, bound to MP and QD, respectively, which provided specific recognition. Second, the introduction of BSA-PEG blocking to reduce nonspecific adsorption. Third, the use of target magnetic separation, which could largely avoid interference from nontarget components in complex matrices. Fourth, the ultrasensitive electrochemical detection of the metal ions obtained by SWV. Finally, the use of SPE for low-cost and portable detection. In a different example, Freitas et al. used MNP to increase the active surface available to immobilize higher biomolecule loads, providing lower LOD and enabling analysis of low (microliter) sample volumes (Freitas et al. 2014). This work was based on the utilization of Ab-modified iron oxide/gold core/shell MNP and QD biolabels for amplified electrochemical immunosensing of Salmonella typhimurium. QD electrochemical detection was accomplished by SWV at C-SPE using mercury(II) nitrate to enhance metal deposition and the stripping performance.