Carbon nanotube composite peptide-based biosensors as putative diagnostic tools for rheumatoid arthritis

Biosensors and Bioelectronics 27 (2011) 113–118

Contents lists available at ScienceDirect

Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios

Carbon nanotube composite peptide-based biosensors as putative diagnostic tools for rheumatoid arthritis María de Gracia Villa a , Cecilia Jiménez-Jorquera a , Isabel Haro b , Maria José Gomara b , Raimon Sanmartí c , César Fernández-Sánchez a,∗ , Ernest Mendoza d,∗ a

Instituto de Microelectrónica de Barcelona, IMB-CNM (CSIC), Campus UAB, 08193 Bellaterra, Barcelona, Spain Unitat de Síntesi i Aplicacions Biomèdiques de Pèptids, Institut de Química Avanc¸ada de Catalunya (IQAC-CSIC), Jordi Girona 18-26, 08034 Barcelona, Spain c Servei de Reumatologia, Hospital Clínic de Barcelona, IDIBAPS, Villarroel 170, 08036 Barcelona, Spain d Grup de Nanomaterials Aplicats, Centre de Recerca en Nanoenginyeria, Universitat Politècnica de Catalunya, c/Pascual i Vila 15, 08028 Barcelona, Spain b

a r t i c l e

i n f o

Article history: Received 18 April 2011 Received in revised form 7 June 2011 Accepted 21 June 2011 Available online 28 June 2011 Keywords: Rheumatoid arthritis Immunosensor Carbon nanotube–polystyrene composite Amperometry

a b s t r a c t This work reports on the fabrication and performance of a simple amperometric immunosensor device to be potentially used for the detection of serum anti-citrullinated peptide antibodies (ACPAs), which are specific for rheumatoid arthritis (RA) autoimmune disease. Sera of RA patients contain antibodies to different citrullinated peptides and proteins such as fibrin or filaggrin. Herein, a chimeric fibrin–filaggrin synthetic peptide (CFFCP1) was used as a recognition element anchored to the surface of a multiwalled carbon nanotube–polystyrene (MWCNT–PS) based electrochemical transducer. The transducer fabrication process is described in detail together with its successful electrochemical performance in terms of repeatability and reproducibility of the corresponding amperometric response. The resulting immunosensor approach was initially tested in sera of rabbits previously inoculated with the synthetic peptide and eventually applied to the detection of ACPAs in human sera. A comparative study was carried out using control serum from a blood donor, which demonstrated the selectivity of the immunosensor response and its sensitivity for the detection of anti-CFFCP1 antibodies present in RA patients. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Rheumatoid arthritis (RA) is a chronic autoimmune disease of unknown origin characterised by chronic inflammation of the joints that can be also followed with systemic complications (Lee and Weinblatt, 2001). RA is a complex multigenic disease in which genetic factors may account for as much as 60% of disease susceptibility (Ollier and Macgregor, 1995; Haro et al., 2011). Anticitrullinated protein/peptide antibodies (ACPAs) are the most specific serological markers for diagnosing RA (Pruijn et al., 2010) and their presence in patient serum is a great help to clinicians in deciding early treatment. Recent studies have indicated that the ACPA response in RA patients is polyclonal and heterogeneous (Verpoort et al., 2006). Antibodies to several citrullinated proteins/peptides such as filaggrin, fibrin or vimentin can be detected in RA patients. Furthermore, in a group of early RA patients, these antibodies were associated with poor radiographic outcome (Sanmarti et al., 2009). Importantly, to prevent disease

∗ Corresponding authors. E-mail addresses: [email protected] (C. Fernández-Sánchez), [email protected] (E. Mendoza). 0956-5663/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.bios.2011.06.026

severity, aggressive therapy given early in the disease course is needed. Therefore, early diagnosis and treatment are required. Considering the heterogeneous response of RA, different analytical tools based on ELISA and microarray technologies, which enabled the reliable detection of this autoimmune disease at an early stage have been reported. Among the former, our group developed a chimeric fibrin/filaggrin synthetic peptide-based ELISA assay highly sensitive and specific for the detection of RA in comparison with healthy controls and those with other chronic diseases (Pérez et al., 2007). Additionally, other ELISA formats were reported and applied to the early detection of RA that make use of different citrullinated protein or peptide receptors (Soós et al., 2007; Enriconi dos Anjos et al., 2009). Autoantigen microarrays containing tens of different peptides were also shown to aid in the diagnosis of RA by being able to detect different autoantibody profiles in patients with RA (Robinson et al., 2002; Hueber et al., 2005; Matsudaira et al., 2008). In spite of the potential of these analytical tools, and considering the incidence of this disease to be between 0.5 and 1% of the population, simpler and more-user friendly devices that could be directly applied in decentralised analysis are required. In this context, a lateral flow immunoassay test for the rapid detection of RA has been developed and commercialised by Euro-Diagnostica AB (www.eurodiagnostica.se). Although this

114

M. de Gracia Villa et al. / Biosensors and Bioelectronics 27 (2011) 113–118

approach enables a rapid screening of the disease, it renders a qualitative visual measurement of serum RA autoantibodies (Liao et al., 2011). Quantitative detection of RA autoantibodies with alternative devices that fulfilled all the above-mentioned requirements would help reduce the number of false positive and negative results and attain a more accurate rapid diagnosis of the disease. Following this technological necessity, the incorporation of nanostructured materials, and in particular multiwalled carbon nanotube (MWCNTs)–polymer composites, to the development of electrochemical immunosensors offers the possibility of developing low cost/high sensitivity devices for point-of-care diagnostics, when specific biomarkers are available as in this particular case (Fernández-Sánchez et al., 2009). In this context, and to the best of our knowledge, only a previous work was reported that described the development of a mass-based biosensor approach for RA detection (Drouvalakis et al., 2008). It relies on the use of a quartz crystal microbalance modified with CNTs, which were further coated with a synthetic peptide receptor. However, its practical use is questionable as it lacks the requirements to be potentially portable and user-friendly. In this work, the development of a simple electrochemical immunosensor composed of a MWCNT–polystyrene (PS) composite transducer and a citrullinated specific peptide receptor is presented. The device performance was assessed using rabbit sera previously immunised with the citrullinated peptide and further tested with human sera of RA patients. Detection limits comparable to those achieved by the previously mentioned ELISA were achieved (Haro et al., 2003). The benefits of this approach are shown in terms of ease of transducer fabrication and covalent modification with the peptide receptor, miniaturization and requirement of a simple and cost-effective instrumentation. The utilisation of this technology in combination with an automated microfluidic system could be applied to the development of automated point-of-care diagnostic devices for low cost immunotesting.

2. Experimental

2.2. Chimeric fibrin–filaggrin peptide synthesis The synthesis of chimeric fibrin–filaggrin synthetic peptide, CFFCP1: [Cit630]␣fibrin (617–631) – S306, S319cyclo[Cys306, 319, Cit312]filaggrin (304–324), is described elsewhere (Pérez et al., 2007). Briefly, it consisted of the synthesis of the chimeric linear sequences in the solid phase and their subsequent cyclisation in solution by means of the formation of a disulfide bridge. The cyclisation was followed by analytical high performance liquid chromatography (HPLC) and the final product purified by semipreparative HPLC. The purified chimeric peptide was successfully characterised by electrospray mass spectrometry. 2.3. Antibody production Anti-CFFCP1 peptide sera were prepared by immunising two rabbits with liposome-entrapped CFFCP1 peptide following the inoculation protocol described in Haro et al. (2003). Liposomes containing this peptide were prepared for the first time following the protocol described in Haro et al. (1995). Briefly, 40 mg of phosphatidylcholine, 20 mg cholesterol and 3 mg of CFFCP1 were evaporated from a chloroform-methanol (2:1, v/v) solution. The samples were submitted to the vacuum of an oil pump to eliminate last traces of the solvents. A 0.9% aqueous NaCl solution was added to the dry lipid-peptide, and the suspension gently stirred for 1 h at room temperature. After preparation, the liposomes were washed two times by successive centrifugation at 25,000 × g and re-suspension of the pellets in 0.9% saline. Two rabbits were immunised subcutaneously with the as-prepared liposomes containing 600 ␮g/mL of CFFCP1 peptide with an equal volume of complete Freund’s adjuvant (Difco, Detroit, MI, USA). Animals were boosted with four injections of 600 ␮g doses of peptide at three-week intervals. Serum samples were collected before each immunisation and also one-week after the last inoculation. Both sera were used as negative and positive rabbit serum samples to test the selectivity and sensitivity of the developed immunosensor, respectively. Animal care procedures were performed in compliance with the Institutional Animal Care and Use Committee Guidelines. The immune reactivity of rabbit sera towards the CFFCP1 peptide was controlled by ELISA as previously described (Haro et al., 2003).

2.1. Fabrication of the composite electrodes 2.4. Human serum specimens The MWCNTs used in the fabrication of the devices are COOHfuncionalised MWCNTs purchased from Nanocyl (NC-31010). The supplier specifications indicate that the MWCNTs are around 10 nm in diameter and according to XPS measurements contain <4% COOH groups. The MWCNT–polystyrene (PS) composite was prepared by dissolving PS pellets (Sigma–Aldrich, Spain) in toluene. Then, MWCNTs were added to the solution and ultrasonicated using a bath sonicator for 3 h. The MWCNT mass percentage in the binary mixture was 10%. The mixture was further homogenized by gentle magnetic stirring, overnight. The composite was deposited on thin film gold or platinum electrodes (100 nm Au/20 nm Cr or 100 nm Pt/20 nm Ni/20 nm Ti), having a geometric area of 4.60 mm2 and fabricated using standard Si/SiO2 /metal microelectronic technology. They were wire-bonded to a PCB stick and packaged using an epoxy resin. Some 4 ␮L of the MWCNT–PS mixture were manually dropped on the surface of the metal electrode using a micropipette, left in air for the solvent to evaporate and then cured for 30 min in an oven at 80 ◦ C. The process was repeated seven times in order to obtain a homogenous composite thick film over the metal electrode. A similar process was carried out on metal electrodes with mixtures containing just PS or MWCNT in order to test the role of both composite components on the electrochemical performance of the resulting composite electrodes.

The serum samples used in this study were from outpatients attending the Rheumatology Unit of the Hospital Clinic of Barcelona. Serum samples used as negative controls were obtained from blood donors at the same hospital. This study was approved by the Ethics Committee of the Hospital Clinic of Barcelona. Samples were firstly tested by ELISA. Here, CFFCP1 was coupled covalently to microtiter plates (Costar Corp., Cambridge, MA), as described previously (Perez et al., 2007). All sera were tested in duplicate. Control sera were also included to monitor inter- and intra-assay variations. Sera were also analysed using second-generation CCP2-based ELISA (Immunoscan, Eurodiagnostica, distributed by Diasorin SA, Madrid, Spain). 2.5. Functionalization of the electrodes and immunoassay protocol Fig. 1 shows a diagram representing the sandwich immunoassay format applied in this study. The diagnostic method is based on the detection of the anti-CFFCP1 antibody RA biomarker. To this purpose, the CFFCP1 peptide was anchored to the surface of the MWCNT electrode. Then, the electrodes were incubated in the sera (human or rabbit). The final step involved the use of anti-human IgG or anti-rabbit IgG secondary antibodies, depending on the sera,

M. de Gracia Villa et al. / Biosensors and Bioelectronics 27 (2011) 113–118

115

Fig. 1. Scheme showing the immunoassay protocol on the MWCNT surface. Chimeric fibrin–filaggrin synthetic peptide bound to the carboxylic groups of the MWCNT interacts with a RA autoantibody, which is detected by interaction with an anti-human antibody conjugated to peroxidase and further enzymatic reaction with TMB substrate.

labelled with horseradish peroxidase (HRP) for electrochemical detection purposes. The CFFCP1 peptide was covalently immobilised on the surface of the carboxylated MWCNTs using the carbodimide/succinimide chemistry (EDC-NHS) (Mendoza et al., 2008). That is, the composite-based electrodes were incubated in a solution containing a 4:1 molar ratio of EDC (1-ethyl3-[3-dimethylaminopropyl]carbodiimide hydrochloride) and N-hydroxysuccinimide-ester (EDC-NHS, Sigma-Aldrich) in deionized water. Following a rinsing step in 10 mM phosphate buffer saline pH 7.2 (PBS), the electrodes were incubated in a 10 ␮g/mL peptide solution in PBS at 4 ◦ C, overnight. Then, they were thoroughly rinsed in PBS containing 0.1% Tween 20 (PBST), in order to remove the non-specifically adsorbed peptide molecules, and further incubated in PBS containing 1% bovine serum albumin (BSA, Sigma–Aldrich) blocking buffer for 1 h at RT. After another rinsing step in PBS, the as fabricated immunosensors were incubated with the human or rabbit sera previously diluted in RIA buffer (10 mM Tris–HCl pH 7.6 containing 1% BSA, 350 mM NaCl, 1% (v/v) Triton X-100, 0.5% (w/v) sodium deoxycholate and 0.1% (w/v) SDS) supplemented with 10% fetal bovine serum (all components from Sigma–Aldrich), for 90 min at room temperature. Following a rinsing step in PBST, the immunosensors were finally incubated in a RIA solution containing 10 ␮g/mL of either goat anti-rabbit IgG or goat anti-human IgG labelled with HRP (Sigma–Aldrich) for 1 h, after which they were thoroughly rinsed in PBST and stored in PBS before carrying out the electrochemical measurement.

Fig. 2. (a) SEM image of the surface of a 10 wt% MWCNT–PS composite. (b) Cyclic voltammogram recorded with a MWCNT–PS composite electrode in a 1.0 mM K3 Fe(CN)6 in 0.1 M KNO3 solution, at 50 mV s−1 . The potential difference between the anodic and cathodic peaks is around 100 mV, which is a similar value to those ones previously reported using other CNT composite materials (Gooding, 2005; Mendoza et al., 2008).

Voltammetric measurements were carried out in a 0.1 M KNO3 solution containing 1.0 mM ferricyanide in order to test the electrochemical performance of the composite electrodes. Studies of the repeatability and reproducibility of the electrode fabrication process were carried out by measuring the amperometric current at a set potential of +0 mV in the same ferricyanide solution. The analytical response of the immunosensor was recorded in a 0.1 M acetate buffer solution pH 5 containing 1 mM H2 O2 enzyme substrate and 0.05 mM 3,3,5,5-tetramethylbenzidine redox mediator (TMB, Sigma–Aldrich). Amperometric measurements were carried out in this solution at a set potential of +150 mV for 200 s. The mean value of the limiting currents recorded in the time interval between 125 s and 175 s was used as the analytical signal. 3. Results and discussion 3.1. Fabrication and performance of MWCNT–PS composite transducers

2.6. Electrochemical measurements Electrochemical experiments were performed at room temperature using a type III ␮-Autolab potentiostat (Ecochemie), controlled with GPES 4.7 (General Purpose Electrochemical System) software package. Measurements were carried out with a standard three-electrode electrochemical cell that comprised a Pt counter electrode, a Ag/AgCl/10% (w/v) KNO3 reference electrode (Orion) and the MWCNT–PS composite working electrode, described above.

MWCNTs were dispersed in PS and the resulting mixture was cast on thin-film metal electrodes followed by drying and curing steps, as described in Section 2. In order to ensure a total coverage of the metal substrate by the MWCNT–PS composite, the above process was repeated several times. Then a uniform film was obtained as can be seen in Fig. 2(A). It appears the MWCNTs are well dispersed within the polymer matrix. The 10% MWCNT content is well above the 3% percolation threshold previously reported for similar composite materials (Dufresne et al., 2002). Other metal electrodes

116

M. de Gracia Villa et al. / Biosensors and Bioelectronics 27 (2011) 113–118

were coated with just PS and MWCNTs in order to test the effect of both components on the performance of the composite electrodes. Fig. 2(b) and Fig. S1 (supplementary information, SI) show the cyclic voltammetric response of the different electrodes in potassium ferricyanide solution. The MWCNT-modified electrode recorded a reversible faradaic signal of the ferrocyanide/ferricyanide redox pair but with a very small faradaic to capacitive ratio due to the extremely high surface active area that the deposited CNT exhibited (Fig. S1(a)). The PS-modified electrode recorded no faradaic signal, this being the expected behaviour considering the non-conducting nature of this polymer (Fig. S1(b)). By contrast, the combination of both MWCNT and PS components gave rise to a composite electrode that recorded a well-defined reversible faradaic signal, similar to that of the metal electrode but with a higher faradaic to capacitive ratio resulting from both a higher faradaic current and a lower capacitive current, which demonstrated the good performance of this material for electrochemical applications. The recorded fast electron transfer kinetics between the composite electrode and solution further confirms that the combined sonication agitation mixing process evenly intercalated the MWCNTs within the polymeric matrix. Using the Randles–Sevcik equation, an estimation of the electrochemical active area was carried out, which was 5.96 mm2 , the roughness factor being 1.29 (estimated considering the geometric area to be the one of the metal electrode). This value is half the one previously reported by our group using a MWCNT–PS composite material containing 30% weight loading of the conductive component (Mendoza et al., 2008). However, the recorded faradaic to capacitive current ratio was again significantly higher using the composite material presented in this work (please, see the low background signal recorded in KNO3 , depicted in Fig. 2(b)). Additionally, although the MWCNTs weight percentage was decreased by 3-fold, the resulting composite surface exhibits similar electron transfer kinetics (compared by the potential difference between the anodic and cathodic peaks of the ferro-/ferricyanide redox pair, which was around 100 mV in both cases). Another added advantage of this composition and preparation approach is that the surface of the MWCNTs is readily exposed to the environment. This is a significant improvement from our previous results, in which an oxygen plasma process was applied to the MWCNT–PS composite surface in order to etch the PS layer that covered the entire composite surface and buried the CNT underneath (Mendoza et al., 2008). Avoiding this activation step simplifies and reduces both the duration and cost of the preparation protocol. The combined sonication agitation effect seems to play a key role in achieving an optimum dispersion of both MWCNT–PS components in the mixture, which directly influenced the conductive properties of the resulting composite material. Nevertheless, it should be mentioned that the plasma activation process added more carboxylic groups to the composite surface on both the MWCNTs and the PS components and so, a higher amount of biological receptors could be attached, which may result in a higher sensitivity of the resulting device (Fernández-Sánchez et al., 2009). Therefore, a compromise between this and adding an extra-step to the preparation of the electrodes, which includes the use of specific instrumentation not readily available in any research lab, should be reached depending on the application to be developed. The reproducibility of the electrode fabrication process is a must condition to develop a reliable diagnostic tool such as the one shown in this work. For this reason, a detailed study of the repeatability and reproducibility of the fabrication process was carried out by measuring the amperometric cathodic current at +0 mV vs. Ag/AgCl reference electrode, recorded with the fabricated electrodes in a 0.1 M KNO3 solution containing 1.0 mM ferricyanide. Fig. 3(a) shows a histogram of the current dispersion recorded with 38 different electrodes fabricated in the same batch using

Fig. 3. (a) Histogram showing the reproducibility of the fabrication process within the same batch, estimated using the amperometric current recorded with 38 electrodes in a 0.1 M KNO3 solution containing 1.0 mM ferricyanide at +0 mV vs. Ag/AgCl ref. electrode. (b) Histogram showing the same parameter for 146 electrodes coming from different batches fabricated over a 3-month period.

the same MWCNT–PS dispersion. The mean current value was −15.5 ± 2.8 ␮A, the coefficient of variation being 18.0%. Furthermore, if the analysis is performed with 146 electrodes coming from different batches fabricated during a 3-month period, the mean current value was −15.6 ± 4.3 ␮A, the coefficient of variation being 27.5% (Fig. 3(b)). These values can be considered to be low taking into account the high degree of manual processing and the different MWCNT–PS dispersions applied to the fabrication of the electrodes. Nevertheless, those electrodes in a batch whose response gave rise to coefficients of variation higher than 10% were discarded. In this context, the batch fabrication yield of those electrodes applied to the development of the immunosensor was estimated to be around 70%. Overall, it can be stated that the fabrication process is quite robust. Work is currently underway to develop a new methodology that avoided manual processing and decreased the variability in the electrochemical response of these composite electrodes. 3.2. Covalent attachment of the peptide receptor to the electrode surface The CFFCP1 peptide was covalently immobilised on the surface of the MWCNTs through an amide bond between amine end groups of the peptide and the carboxylic groups of the CNTs. Considering the peptide structure and the pH at which the immobilisation was carried out (pH = 7.2), it is likely the peptide interacted

M. de Gracia Villa et al. / Biosensors and Bioelectronics 27 (2011) 113–118

117

with the carboxylated CNTs through the ␣-amino group of the Nterminal aminoacid (pKa = 8.9) or the ␧-amino group of the two lysine residues (pKa = 10.5) of its structure. Therefore, three main possible orientations of the immobilised peptide receptor on the transducer surface could be attained. Fig. S2 in SI shows the peptide linear structure and the peptide 3D-structure of minimum energy, where these amino residues are identified.

3.3. Amperometric response of the peptide-based biosensor in rabbit serum samples The amperometric response of the immunosensor was recorded using TMB as redox mediator of the HRP reaction. HRP catalyses the reduction of H2 O2 taking place with the concomitant oxidation of TMB to TMB2+ . This oxidised species could be reduced back to TMB substrate by applying a set overpotential of +150 mV vs. an Ag/AgCl reference electrode, as described elsewhere (Mendoza et al., 2008; Fernández-Sánchez et al., 2009). Initial measurements with the device were carried out with positive and negative rabbit sera diluted 400-fold in PBS and RIA buffers, in order to account for the potential contribution of the non-specific interactions of both serum and antibody conjugate with the immunosensor to the analytical signal. Fig. S3 in SI shows a bar graph representing the currents recorded for each experiment. The signal for the positive rabbit serum was 30-fold and 9-fold higher than that for the negative serum when RIA and PBS buffers were used, respectively. Therefore, it was shown that the RIA buffer was highly convenient in order to minimise such non-specific interactions and thus get a more specific response with the developed immunosensor. The same behaviour was previously reported in the development of ELISA assays for the detection of RA autoantibodies using the same peptide (Pérez et al., 2007). Fig. 4(a) shows the amperometric responses recorded with different immunosensors for eight serial dilutions of a positive rabbit serum and the response recorded with one dilution of the negative rabbit serum, for comparative purposes. The immunosensor response is inversely proportional to the positive serum dilution and the resulting calibration plot is depicted in Fig. 4(b), together with the responses for the same dilutions of the negative serum. The amperometric current values recorded with both sera are clearly different for dilutions down to 1/6400 where the values start to overlap. These results clearly show that the device selectively and sensitively responded to the anti-CFFCP1 antibodies presented in the positive serum.

Fig. 4. (a) Amperometric responses to increasing dilutions of rabbit positive serum ranging from 1/200 to 1/25,600 prepared in RIA buffer (the arrow indicate how the signal decreases for increasing dilutions of the serum). Detection potential: +150 mV vs. Ag/AgCl reference electrode. The amperometric signals recorded without serum and with a 1/800 dilution of rabbit negative serum are overlapped with those signals corresponding to the higher rabbit positive serum dilutions. Two measurements for each dilution were carried out using two different immunosensors. One measurement for each dilution is shown for clarity purposes. (b) Corresponding calibration plots for rabbit positive and negative sera. Each point corresponds to the mean value of the two measurements for each serum dilution, the error bars representing the corresponding standard deviation. Data was fitted to a second order polynomial equation without considering the error bars weight.

3.4. Analysis of human serum samples The assessment of the performance of the immunosensor as a potential tool for the rapid diagnosis of rheumatoid arthritis disease was carried out using human serum samples of two patients suffering from RA and a healthy blood donor. Two dilutions of each of these sera were tested. Fig. 5 shows a bar diagram depicting the immunosensor responses for each dilution. A difference between the responses of the immunosensors in both RA established serum and the control (blood donor) is clearly observed for both tested dilutions. The response to the borderline serum is slightly higher than that to the control one but considering the error bars of both samples the current differences are not significant. Nevertheless, it should be stressed that both RA patient samples were chosen from the values previously obtained by ELISA, which were representative of the lower and higher RA autoantibody concentration ranges in serum. Work is in progress to assess the immunosensor performance with other serum samples and to improve its response in terms of repeatability and reproducibility.

Fig. 5. Bar plots showing the amperometric immunosensor responses to two different dilutions of human sera coming from two patients suffering from RA and a blood donor.

118

M. de Gracia Villa et al. / Biosensors and Bioelectronics 27 (2011) 113–118

4. Conclusion

References

MWCNT–PS composite based electrodes were applied to the development of an electrochemical immunosensor device for the rapid diagnosis of RA. The fabrication of the electrodes proved to be repetitive and reproducible from the electrochemical point of view, although the fabrication process, and in particular the composite deposition step, is still manual. In this sense, we are now developing a process to perform an automated deposition of the composite and reduce even further the electrode fabrication cost and the dispersion of its electrochemical response. The CNTs were covalently functionalized with the peptide CFFCP1. This synthetic peptide was previously shown to be specific against autoantibodies present in RA patients, and therefore could be used to diagnose this particular disease. The biosensor was tested with both sera from immunised rabbits as well as sera from human patients. In all the cases, the technology successfully performed for the detection of RA autoantibodies, although a more in depth assessment of its potential should be carried out working with a large number of patient samples, which solidly demonstrated the viability of this device for the development of rapid and economic point-of-care diagnostic tools. In this regard, future development must also include the integration with a microfluidic system to enable automated operation.

Dufresne, A., Paillet, M., Putaux, J.L., et al., 2002. Journal of Materials Science 37 (18), 3915–3923. Drouvalakis, K.A., Bangsaruntip, S., Hueber, W., et al., 2008. Biosensors & Bioelectronics 23 (10), 1413–1421. Enriconi dos Anjos, L.M., Pereira, I.A., d’Orsi, E., et al., 2009. Clinical Rheumatology 28 (2), 153–158. Fernández-Sánchez, C., Pellicer, E., Orozco, J., et al., 2009. Nanotechnology 20 (33), 335501. Gooding, J.J., 2005. Electrochimica Acta 50 (15), 3049–3060. Haro, I., Gomara, M.J., Pérez, M.L., et al., 2011. European Journal of Medicinal Chemistry 46 (4), 1095–1102. Haro, I., Perez, S., Garcia, M., Chan, W.C., Ercilla, G., 2003. FEBS Letters 540 (1–3), 133–140. Haro, I., Pinto, R.M., Gonzalez, et al., 1995. Microbiology and Immunology 39 (7), 485–490. Hueber, W., Kidd, B.A., Tomooka, et al., 2005. Arthritis & Rheumatism 52 (9), 2645–2655. Lee, D.M., Weinblatt, M.E., 2001. Lancet 358 (9285), 903–911. Liao, J., Ip, W.-S., Cheung, K.-Y., et al., 2011. Clinica Chimica Acta 412 (9–10), 778–781. Matsudaira, T., Tsuzuki, S., Wada, A., et al., 2008. Biotechnology Progress 24, 1384–1392. Mendoza, E., Orozco, J., Jiménez-Jorquera, C., et al., 2008. Nanotechnology 19 (7), 075102. Ollier, W.E.R., Macgregor, A., 1995. British Medical Bulletin 51 (2), 267–285. Pérez, M.L., Gomara, M.J., Ercilla, G., et al., 2007. Journal of Medicinal Chemistry 50 (15), 3573–3584. Pruijn, G.J.M., Wiik, A., Venrooij, W.J. van, 2010. Arthritis Research & Therapy 12 (1), 203. Robinson, W.H., DiGennaro, C., Hueber, W., et al., 2002. Nature Medicine 8 (3), 295–301. Sanmarti, R., Graell, E., Perez, M.L., et al., 2009. Arthritis Research & Therapy 11 (5), R135. Soós, L., Szekanecz, Z., Szabó, Z., et al., 2007. Journal of Rheumatology 34 (8), 1658–1663. Verpoort, K.N., Jol-van der Zijde, C.M., Papendrecht-van der Voort, E.A., et al., 2006. Arthritis and Rheumatism 54 (12), 3799–3808.

Acknowledgements Partial financial support from Consolider-Ingenio 2010 project, ref. CSD2006-00012 and project ref. CTQ2009-13969-C02-01/BQU, both from MICINN is acknowledged. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.bios.2011.06.026.