Article pubs.acs.org/ac
Ultrasensitive Label-Free Immunoassay for Optical Determination of Amitriptyline and Related Tricyclic Antidepressants in Human Serum Anne Katrin Krieg* and Günter Gauglitz Institute of Physical and Theoretical Chemistry, Eberhard Karls University, Auf der Morgenstelle 18, 72076 Tuebingen, Germany ABSTRACT: The present work focuses on the development of a label-free and ultrasensitive immunoassay for the detection of the drug amitriptyline in human serum. Reflectometric interference spectroscopy is used as the detection method, providing a simple, but highly sensitive optical setup. Amitriptyline is a common antidepressant; however, it has a small therapeutic window and can cause severe side effects in case of wrong dosage. Therefore, it is highly recommended for therapeutic drug monitoring to control the drug level. The limit of detection for this optical immunosensor was determined in buffer (0.3 μg/L) and in human serum (0.5 μg/L). It has become evident that this assay can compete with HPLC measurements. For drug concentrations at a normal level or above, the sample can be diluted up to 1:100. Especially for limited sample volumes, this is a great advantage. The sensor surface shows very high stability, and together with the regeneration solution 80 measurement cycles can be performed on each transducer chip. Cross-reactivity experiments indicate that a sum determination of several tricyclic antidepressants is possible.
A
example, AMT, is monitored at the beginning of the medication therapy to set the patient’s dose individually. In case of changed physiological conditions, additional drug administration or suspected drug levels outside the therapeutic window, the concentration levels are monitored also during the medication therapy. The therapeutic concentration range is 120−250 mg/L AMT, with toxic levels starting at 500 mg/L AMT.11 Another application area to be addressed is the forensic toxicology. In this field, limited sample volume, complex sample matrices, and determination of trace concentrations make it challenging for analysts to detect all substances of interest. Several forensic studies demonstrated that amitriptyline can be analyzed even in bones and bone marrow.12−15 Biosensors are already used in medical diagnostics, and the number of available tests with biological recognition elements for various clinical parameters is increasing.16−18 In particular, immunoassays are used because of high antibody specificity, low limits of detection and the possibility to measure in complex matrices (e.g., serum, blood, saliva, urine). The present work is based on a heterogeneous immunoassay format using the optical detection method reflectometric interference spectroscopy (RIfS). This label-free method allows the time-resolved observation of biomolecular binding interactions on the sensor surface.3,19 Various label-free sensors can be found in literature for different application fields.20−25
ccording to the WHO, 350 million people were affected by depressive disorders in 2012.1 The global burden of disease (GBD) study in 2008 revealed the unipolar depression (major depression) as the third leading cause of disability worldwide, and in 2030 it will be in the lead. Especially for women aged 15−44, mental disorders are the main reason for so-called lost years of a healthy life.2,3 Tricyclic antidepressants (TCA), and its representative amitriptyline (AMT), have been well-known and well proven for a long time in treatment of depressions. Because of this long experience and their effectiveness, they are still readily prescribed.4 Nevertheless, TCAs show severe side effects (e.g., cardiotoxicity) because of the general reuptake inhibition of the neurotransmitters serotonin, dopamine and norepinephrine.5 Together with international pharmacology societies the national German consortium “Arbeitsgemeinschaft für Neuropsychopharmakologie und Pharmakopsychiatrie” (AGNP) highly recommends Therapeutic Drug Monitoring (TDM) for TCAs, including amitriptyline and nortriptyline (NRT).6 Depending on genetic predisposition or physiological conditions, the speed of drug transformation can vary significantly. Up to a 20-fold interindividual variation at the same dose of medication was observed.6,7 This can lead to a lack of therapeutic effect or to severe overdose side-effects. Unfortunately, TDM is used in practice only rarely,8 although the positive effects have been shown in numerous studies.9,10 Therefore, TDM can improve the therapeutic drug management in case of questionable compliance, suspected toxicity or drug−drug interaction.11 The concentration level of a drug, for © XXXX American Chemical Society
Received: May 21, 2015 Accepted: July 24, 2015
A
DOI: 10.1021/acs.analchem.5b01895 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
succinimide (NHS) as coupling reagents. NHS (5.7 mg) and DIC (12 μL) are dissolved in 36 μL of DMF. On each transducer, 8 μL are applied, and the transducer sandwich is stored in a DMF-vapor saturated chamber for 4 h. After the transducers were washed with DMF and acetone and dried under nitrogen, the transducers are wet out with 15 μL NRTsolution (2 mg/mL in milli-Q water). The transducer sandwich is stored in a water-vapor saturated chamber for 1−2 days. Prior to the measurement, the transducers are washed with milli-Q water and dried under nitrogen. Reflectometric Interference Spectroscopy. For quantitative detection of TCAs, the optical technique reflectometric interference spectroscopy (RIfS) was used. RIfS has proven to be a very suitable method for sensitive and quantitative chemoand biosensors in various fields of applications (e.g., medicine, food).3,19,22,32,33 This technique is based on interference of white light at thin films. A change in optical thickness (the product of physical thickness and refractive index) causes a shift of the interference spectrum which can be detected in a timeresolved manner.34 More detailed information about the method and the setup can be found in literature.35−37 Assay Procedure. RIfS is a label-free method, therefore, no labeling procedures are needed. As the analyte AMT is a small molecule (277 Da), only direct or competitive assay formats are possible. In our case, the binding-inhibition test was the only appropriate format. An overview of the measuring procedure is shown in Figure 1. The transducer surface is coated with the
Currently, amitriptyline is measured in clinical laboratories with HPLC/MS especially for overdose concentrations.4,5,26−28 This technique requires expensive instrumentation and trained staff. Furthermore, not all laboratories can measure serum samples without prior liquid/liquid extraction. For immunoassays, on the other hand, simple, small, and therefore, cheaper instrumentation is sufficient, and no trained personnel or sample pretreatment is required. Recent immunoassays show a lack of sensitivity and are not applicable on trace level determinations.29−31 This newly developed ultrasensitive immunoassay has a much improved LOD and can be compared with HPLC measurements. As a matter of principal, antibodies against small drug molecules suffer from reduced selectivity toward molecules with similar structure. This fact can become an advantage when using it for sum determination of substances with similar structure and pharmacological effect. Therefore, the cross-reactivity of the monoclonal anti-AMT antibody was tested. These experiments revealed the structure moieties of the drug molecule to be mandatory for the interaction with the antibody used.
■
EXPERIMENTAL SECTION Reagents and Materials. RIfS glass transducers (1 cm × 1 cm) having a 1 mm-D263-glass substrate with a layer of 10 nm Ta2O5 covered with 330 nm SiO2 on top are obtained from Schott AG, Mainz, Germany. Bis-amino-poly(ethylenglycol) (DA-PEG; 2000 da) is purchased from Rapp Polymere Tuebingen, Germany. All tricyclic antidepressants and common chemicals of analytical grade are purchased from Sigma-Aldrich, Deisenhof, Germany, and from Merck KGaA, Darmstadt, Germany. 3-Glycidyloxypropyl-trimethoxysilane (GOPTS) is purchased from Fluka, Neu-Ulm, Germany. The human serum, the lyophilized protein powders ovalbumin (OVA), and bovine serum albumin (BSA) are purchased from Sigma-Aldrich, Deisenhof, Germany. Monoclonal mouse antibodies to amitriptyline are purchased from Acris Antibodies GmbH, Herford, Germany (abI, Clone 202), and from abcam plc, Cambridge, United Kingdom (abII, Clone 8.F.9). For regeneration of the sensor surface, a solution of guanidine hydrochloride (GdnHCl) (6 M, pH 1.5) in milli-Q water is used. As buffer solution, phosphate-buffered saline (PBS) was prepared using 150 mM NaCl and 10 mM KH2PO4, and the pH was adjusted to 7.4. Preparation of the Sensor Surface. The SiO2 surface of the transducer is cleaned and activated using freshly prepared Piranha solution (mixture of conc. H2SO4 and 30% H2O2; ratio 3 to 2). Next, the glass surface is modified with 3Glycidyloxypropyl-trimethoxysilane (GOPTS). Each transducer is wet out with 15 μL GOPTS and covered with another transducer for 1 h. This sandwich technique is used in each of the following surface chemistry steps. The transducers are cleaned with acetone and dried under nitrogen. Afterward, the biopolymer aminodextrane (AMD) is bound covalently onto the GOPTS layer using 10 μL of dissolved AMD (1 mg/mL in milli-Q water). After 1 day in a water−vapor saturated chamber, the transducers are cleaned with water and dried under nitrogen. The amino functions of AMD are transferred into carboxyl-functions using 10 μL of dissolved glutaric acid (2 mg/ μL glutaric acid in DMF). After at least 6 h, the transducers are washed with DMF and water and dried under nitrogen. Subsequently, nortriptyline (NRT) as demethylated derivative of amitriptyline (AMT) is immobilized on the sensor surface using N,N′-diisopropyl-carbodiimide (DIC) and N-hydroxy-
Figure 1. Antigen molecules are immobilized on the sensor surface. Example measurement illustrates the procedure of a binding-inhibition assay: (1) buffer, (2) preincubated sample, (3) flushing with buffer, (4) removing antibodies by regeneration solution, and (5) buffer.
AMT derivative NRT as described in the section before. Because a binding-inhibition test format is used each sample measurement has to be related to a blank measurement. With this measurement the maximum binding signal of a defined amount of antibodies is determined. Furthermore, prior to the sample measurements, there is an incubation step. The sample containing an unknown AMT concentration is incubated with the same amount of anti-AMT antibodies as for the blank measurement (in the range of 1−3 mg/mL) for 30 min at room temperature. Antibodies are more stable at high protein concentrations. Therefore, all antibody solutions were stabilized with 1 mg/mL OVA or 1 mg/mL BSA. The transducer surface is flushed with buffer (1). The sample incubated with a defined amount of antibodies is brought to the B
DOI: 10.1021/acs.analchem.5b01895 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry surface (2) showing the mass transport limited linear binding signal. Afterward, the flow cell is flushed again with buffer, leaving only those antibodies behind that were not blocked by AMT (sample) during incubation time (3). To regenerate the surface, guanidine hydrochloride (GdnHCl, 6 M, pH 1.5) is used, removing the antibodies due to degeneration (4). Flushing the flow cell with buffer again recovers the transducer surface (5), and the next measurement starts again with step 1. Immunoassay Evaluation. As the binding of the antibodies is limited by mass transport, the binding signals show a linear and concentration-dependent slope during sample injection.38−41 A linear regression is performed, and each slope is normalized to the mean of the blank measurements (zero AMT concentration). The mathematical relationship between relative slope and AMT concentration can be identified by the use of a logistic fit function with four parameters eq 1. A1 is the upper and A2 the lower asymptote for the binding inhibition assay. The analyte concentration, with a signal decrease of 50% of the dynamic signal range (IC50), is defined as the inflection point x0. The parameter p indicates the slope of the tangent in this point.42 y = A2 +
the presented immunoassay are very low and can easily compete with various other standard techniques (e.g., UHPLCMS/MS, LC-MS/MS, GC, UV).4,26,27,49,50 To prove the validity of our calibration function, the recovery rates of three different concentrations at 10%, 50%, and 90% of the dynamic signal are performed. For 0.7, 4.0, and 12.0 μg/L AMT, the recovery rates are calculated to be 96.9 ± 9.9%, 103.4 ± 3.4%, and 96.3 ± 0.8%, respectively. This confirms again the very good performance of this immunoassay. Transducer Performance. The intertransducer reproducibility and the long-term stability of six transducers prepared in parallel are tested (see Figure 3) to verify the applicability of
A1 − A 2 1+
y = A2 +
Figure 2. Calibration and performance of the AMT assay in PBS: fiveparameter logistic fit (red) and 95% confidence band (pink); 1.9 μg/L anti-AMT (abI); R2 = 0.999, χ2 = 1.101.
() x x0
p
(1)
A1 − A 2 ⎡ 1 + (21/ s − 1) ⎣⎢
p ⎤s
( ) ⎦⎥ x x0
(2)
Some antigen−antibody systems show a more asymmetric sigmoid curve shape for their concentration dependency. In these cases, the logistic fit function requires a fifth parameter eq 2 for a better matching with the measured values. The additional parameter s influences the symmetry of the curve.43
■
RESULTS AND DISCUSSION Assay Performance in Buffer. Label-free immunoassays often suffer from unspecific binding by matrix components. For a better shielding, the transducer surface is coated with biopolymers, for example, functionalized dextran or polyethylene glycol.44,45 In preliminary works, the two biopolymer coatings amino-dextran (AMD)46 and diamino-polyethylene glycol (DA-PEG)47 were tested and compared.48 Even though contact angle measurements revealed a higher NRT surface loading for AMD-coated transducers, DA-PEG-coated transducers show higher and foremost mass-transport-limited binding signals. This indicates a better accessibility of the antigens on DA-PEG surfaces. A suitable regeneration solution was found in previous studies, too.48 After characterization and optimization of the heterogeneous antibody−antigen interaction, a concentration series was performed. Binding signals of 1.9 μg/L anti-AMT antibodies (abI) inhibited with ten different AMT concentrations were determined as 3-fold measurement (see Figure 2). A calibration function is calculated using the five-parameter logistic fit function (see Immunoassay Evaluation section). Because of the very low standard deviations, the confidence band (95%) is very narrow, demonstrating the exceptionally high reproducibility of this biosensor. The LOD and LOQ are calculated to be 268 and 829 ng/L AMT, respectively. MDC and RDL are more common in medical analytics and are calculated to be 160 and 262 ng/L AMT, respectively. Altogether, the detection limits of
Figure 3. Intertransducer stability (gray bars) and long-term-stability over 10 weeks (hatched bars); six transducers related to three sandwiches; position during preparation: top (t), down (d).
the sensor coating. Therefore, on each transducer the binding signal of 1.2 mg/L anti-AMT-antibodies (abII) was determined in a 3-fold measurement directly after the sensor preparation. The relative variation of the mean value of the binding signal amounts to ±12.6% with standard deviations in the range of 0.41−2.97% (see Figure 3, gray bars). After 10 weeks (74 days) of dry storage at 3 °C in the refrigerator, the same transducers were tested again in the same manner, and the binding signals were compared. The signal slope is reduced by 9−39% (see Figure 3, hatched bars). However, this reduction does not affect the immunoassay so much as the assay format is a binding inhibition test. Before each sample measurement, a blank measurement has to be carried out determining the maximum of the binding signal or signal slope. All of the following measurements are related to C
DOI: 10.1021/acs.analchem.5b01895 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry
(NRT, OPP) also do not affect the binding sites very much. Such studies can help to examine which structural unit of the molecule is important for binding with a specific antibody. In case of AMT, it is the tricyclic structure. Assay Performance in Human Serum. AMT is usually quantified in human plasma or serum. Therefore, the assay performance is tested in 1:10 diluted human serum spiked with seven different AMT concentrations. Triple measurements were carried out and calibrated (Figure 6) using a five-
this value. Altogether, the sensor coating showed a high stability and reproducibility. Concerning long-term stability, this assay allows up to 80 measurements on each transducer using GdnHCl as regeneration solution. The binding signal slope is still at 99.2% after this series of measurements. Cross-Reactivity of the Antibody. It is well-known that antibodies against small molecules show high cross-reactivity toward structurally similar compounds compared to antiprotein antibodies.11,51−53 Therefore, the ability to inhibit the binding of anti-AMT-antibody to AMT was tested and quantitatively compared with six different TCAs and derivatives respectively (see Figure 4).
Figure 6. Calibration (red) and performance of the AMT assay in human serum matrix; five-parameter logistic fit (red) and 95% confidence band (pink); 2.1 μg/L anti-AMT (abII); R2 = 0.998, χ2 = 0.613.
parameter fit function according to Figure 2. The detection limits calculated are the limit of detection (LOD = 540 ng/L), limit of quantification (LOQ = 1780 ng/L), minimum detectable concentration (MDC = 230 ng/L), and the reliable detection limit (RDL = 530 ng/L). Taking into account the dilution factor of 10, the LOD and LOQ for real samples are 5.4 and 17.8 μg/L AMT, respectively. To verify the measurement results, four spiked AMT concentrations in human serum are determined by the new immunoassay and by a validated HPLC-MS setup. The correlation of both methods is shown in Figure 7. Furthermore, recovery rates in human serum for both methods were determined and compared (Table 1). The recovery rates for the immunoassay using RIfS are in the range of 73.6−103.4%, whereas the HPLC-MS recovery range is 71.7−163.0%. It should be noted that the HPLC-MS method was validated only for concentrations greater than 30 μg/L AMT and only double determination was performed.
Figure 4. Chemical structures of amitriptyline and six structurally related TCAs.
This evaluation revealed certain positions, where slight structural differences affect the binding interaction significantly. The ability of 10 nmol TCA blocking the anti-AMT-antibody binding sites was analyzed (see Figure 5). The addition of a big
■
CONCLUSION The high reproducibility and robustness of the sensor coating together with the assay format indicates an excellent reliability. This immunoassay for AMT shows an extremely low limit of detection with 268 ng/L AMT in buffer, as well as 540 ng/L AMT in diluted human serum (1:10), comparable or better than chromatographic methods such as HPLC. The reliability in real matrix was also validated with HPLC-MS measurements. Furthermore, with this immunoassay it is possible to determine the sum of structurally similar and active TCAs as well as their metabolites. The RIfS method allows small and cost-effective instrumentation. In conclusion, an ultrasensitive immunoassay was developed for the use in therapeutic drug monitoring (high concen-
Figure 5. Ability of 10 nmol AMT and six other TCAs (see Figure 3) to inhibit 1.53 mg/L anti-AMT-antibodies (abI).
chlorine atom at the tricyclic structural unit caused the highest effect (CLOM). Also, an additional methyl side chain close to the tricyclic structure (TRIM) led to a considerably reduction of the binding inhibition ability. Only small decreases were observed using derivatives with carbon atoms substituted with oxygen or nitrogen within the tricyclic structure (IMI, DOX, OPP). Furthermore, changes at the end of the side chain D
DOI: 10.1021/acs.analchem.5b01895 Anal. Chem. XXXX, XXX, XXX−XXX
Analytical Chemistry
Article
■
ACKNOWLEDGMENTS We thank the laboratory of the “Klinikum rechts der Isar” in Munich for the HPLC-MS measurements of AMT spiked human serum.
■
(1) World Health Organization. Depression is a common illness and people suffering from depression need support and treatment: WHO marks 20th Anniversary of World Mental Health Day, 2012. http:// www.who.int/mediacentre/news/notes/2012/mental_health_day_ 20121009/en/ (2) World Health Organization. The global burden of disease2004 update, 2008. http://www.who.int/healthinfo/global_burden_ disease/2004_report_update/en/. (3) Krieg, A. K.; Hess, S.; Gauglitz, G. Proc. SPIE 2013, 8774, 87740H. (4) Poklis, J. L.; Wolf, C. E.; Goldstein, A.; Wolfe, M. L.; Poklis, A. J. Clin. Lab. Anal. 2012, 26, 286−294. (5) Khatoon, A.; Setapar, S.; Ahmad, A.; Nian-Yian, L. Pharma Chem. 2013, 5, 79−85. (6) Hiemke, C.; Baumann, P.; Bergemann, N.; Conca, A.; Dietmaier, O.; Egberts, K.; Fric, M.; Gerlach, M.; Greiner, C.; Gründer, G.; Haen, E.; Havemann-Reinecke, U.; Jaquenoud Sirot, E.; Kirchherr, H.; Laux, G.; Lutz, U. C.; Messer, T.; Müller, M. J.; Pfuhlmann, B.; Rambeck, B.; Riederer, P.; Schoppek, B.; Stingl, J.; Uhr, M.; Ulrich, S.; Waschgler, R.; Zernig, G. Pharmacopsychiatry 2011, 44, 195−235. (7) Burke, M. J.; Preskorn, S. H. In Antidepressants: Past, Present and Future; Preskorn, S. H., Feighner, J. P., Stanga, C. Y., Ross, R., Eds.; Springer: Berlin, 2004; pp 87−114. (8) Eliasson, E.; Lindh, J.; Malmström, R.; Beck, O.; Dahl, M.-L. Eur. J. Clin. Pharmacol. 2013, 69, 25−32. (9) Choong, E.; Rudaz, S.; Kottelat, A.; Guillarme, D.; Veuthey, J.-L.; Eap, C. B. J. Pharm. Biomed. Anal. 2009, 50, 1000−1008. (10) Wille, S. M. R.; Cooreman, S. G.; Neels, H. M.; Lambert, W. E. E. Crit. Rev. Clin. Lab. Sci. 2008, 45, 25−89. (11) Linder, M. W.; Keck, P. E. Clin. Chem. 1998, 44, 1073−1084. (12) Watterson, J. H.; Cornthwaite, H. M. J. Anal. Toxicol. 2013, 37 (8), 565−572. (13) Desrosiers, N. A.; Watterson, J. H.; Dean, D.; Wyman, J. F. J. Forensic Sci. 2012, 57, 544−549. (14) McGrath, K. K.; Jenkins, A. J. Am. J. Forensic. Med. Pathol. 2009, 30, 40−44. (15) Drummer, O. In Drug Testing in Alternate Biological Specimens; Jenkins, A.; Caplan, Y., Eds.; Humana Press: Totowa, NJ, 2008; pp 131−137. (16) Bottazzi, B.; Fornasari, L.; Frangolho, A.; Giudicatti, S.; Mantovani, A.; Marabelli, F.; Marchesini, G.; Pellacani, P.; Therisod, R.; Valsesia, A. J. Biomed. Opt. 2014, 19, 017006. (17) Orgovan, N.; Salánki, R.; Sándor, N.; Bajtay, Z.; Erdei, A.; Szabó, B.; Horvath, R. Biosens. Bioelectron. 2014, 54, 339−344. (18) Kumar, S.; Kumar, S.; Ali, M. A.; Anand, P.; Agrawal, V. V.; John, R.; Maji, S.; Malhotra, B. D. Biotechnol. J. 2013, 8, 1267−1279. (19) Rau, S.; Gauglitz, G. Anal. Bioanal. Chem. 2012, 402, 529−536. (20) Bertók, T.; Katrlík, J.; Gemeiner, P.; Tkac, J. Microchim. Acta 2013, 180, 1−13. (21) Sun, Y.-S. Instrum. Sci. Technol. 2014, 42, 109−127. (22) Ewald, M.; Le Blanc, A. F.; Gauglitz, G.; Proll, G. Anal. Bioanal. Chem. 2013, 405, 6461−6469. (23) Gauglitz, G. Proc. SPIE 2013, 8774, 87740I. (24) Kolarov, F.; Niedergall, K.; Bach, M.; Tovar, G. M.; Gauglitz, G. Anal. Bioanal. Chem. 2012, 402, 3245−3252. (25) Fang, Y. Int. J. Electrochem. 2011, 2011, 460850. (26) Sarafraz Yazdi, A.; Razavi, N. Chromatographia 2011, 73, 549− 557. (27) Sheng, J.; Lei, J.; Ju, H.; Song, C.; Zhang, D. Anal. Chim. Acta 2010, 679, 1−6. (28) Tybring, G.; Nordin, J.; Widén, J. J. Chromatogr., Biomed. Appl. 1998, 716, 382−386.
Figure 7. Correlation of RIfS and HPLC-MS measurements with spiked human serum samples (according to Table 1); correlation slope = 0.92 with R2 = 0.994; dashed line shows slope = 1.
Table 1. Recovery Rates in Human Serum (Pooled), Spiked with Four AMT Concentrations (Stock Solution 5, 15, 50, 100 μg/L), Measured (a) with the RIfS Setup (Dilution 1:10, Triple Measurement) and (b) with Validateda HPLC-MS (Pure, Double Determination) measured conc. [μg/L] spiked conc. [μg/L]
a
b
5 15 50 100
3.7 15.5 44.0 75.0
8.2 15.2 38.9 69.9
a
recovery rates [%] a 73.6 103.4 88.0 75.0
± ± ± ±
b 9.7 10.3 1.8 2.4
163.0 100.3 71.7 69.9
± ± ± ±
0.3 0.6 9.1a 1.3a
HPLC-MS setup was validated for concentrations >30 μg/L.
trations, sum determination) or forensic investigations (low concentrations, effect based determination) of TCAs. Nevertheless, the selectivity of antibodies against small molecules should be further increased to be able to use this intrinsic and exceptional advantage of antibodies for quantification of single target molecules. There are several studies demonstrating an increase of selectivity because of fragmentation of the antibodies. Some Fab and (Fab)2 fragments show higher selectivity and stability compare to the whole IgG antibody. This can be promising way for further selectivity improvements of antibodies against small drug molecules.
■
REFERENCES
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Phone: +49-70712972651. Fax: +49-7071-295772. Author Contributions
The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes
The authors declare no competing financial interest. E
DOI: 10.1021/acs.analchem.5b01895 Anal. Chem. XXXX, XXX, XXX−XXX
Article
Analytical Chemistry (29) Melanson, S. E. F.; Lewandrowski, E. L.; Griggs, D. A.; Flood, J. G. J. Anal. Toxicol. 2007, 31, 270−275. (30) Meenan, G. M.; Barlotta, S.; Lehrer, M. J. Anal. Toxicol. 1990, 14, 273−276. (31) Poklis, A.; Soghoian, D.; Richard Crooks, C.; Saady, J. J. Clin. Toxicol. 1990, 28, 235−248. (32) Pröll, F.; Möhrle, B.; Kumpf, M.; Gauglitz, G. Anal. Bioanal. Chem. 2005, 382, 1889−1894. (33) Bleher, O.; Ehni, M.; Gauglitz, G. Anal. Bioanal. Chem. 2012, 402, 349−356. (34) Proll, G.; Markovic, G.; Steinle, L.; Gauglitz, G. In Biosensors and Biodetection; Rasooly, A., Herold, K., Eds.; Humana Press: New York, 2009; pp 167−178. (35) Gauglitz, G.; Brecht, A.; Kraus, G.; Nahm, W. Sens. Actuators, B 1993, 11, 21−27. (36) Schmitt, H.-M.; Brecht, A.; Piehler, J.; Gauglitz, G. Biosens. Bioelectron. 1997, 12, 809−816. (37) Gauglitz, G.; Proll, G. In Biosensing for the 21st Century; Renneberg, R., Lisdat, F., Eds.; Springer: Berlin, 2008; pp 395−432. (38) Glaser, R. W. Anal. Biochem. 1993, 213, 152−161. (39) Schuck, P.; Minton, A. P. Anal. Biochem. 1996, 240, 262−272. (40) Richalet-Sécordel, P. M.; Rauffer-Bruyère, N.; Christensen, L. L. H.; Ofenloch-Haehnle, B.; Seidel, C.; Van Regenmortel, M. H. V. Anal. Biochem. 1997, 249, 165−173. (41) Wang, G.; Driskell, J. D.; Porter, M. D.; Lipert, R. J. Anal. Chem. 2009, 81, 6175−6185. (42) Dudley, R. A.; Edwards, P.; Ekins, R. P.; Finney, D. J.; McKenzie, I. G.; Raab, G. M.; Rodbard, D.; Rodgers, R. P. Clin. Chem. 1985, 31, 1264−1271. (43) Liao, J. J. Z.; Liu, R. J. Chemom. 2009, 23, 248−253. (44) Sämann, M.; Furin, D.; Thielmann, J.; Pfäfflin, A.; Proll, G.; Harendt, C.; Gauglitz, G.; Schleicher, E.; Schubert, M. B. Phys. Status Solidi C 2010, 7, 1160−1163. (45) Schwarz, B.; Schweizer, N.; Pörll, F.; Proll, G.; Gauglitz, G. Procedia Eng. 2010, 5, 906−909. (46) Piehler, J.; Brecht, A.; Valiokas, R.; Liedberg, B.; Gauglitz, G. Biosens. Bioelectron. 2000, 15, 473−481. (47) Piehler, J.; Brecht, A.; Geckeler, K. E.; Gauglitz, G. Biosens. Bioelectron. 1996, 11, 579−590. (48) Krieg, A. K.; Hess, S.; Gauglitz, G. Proc. SPIE 8774 2013, 87740H. (49) Remane, D.; Meyer, M.; Wissenbach, D.; Maurer, H. Anal. Bioanal. Chem. 2011, 400, 2093−2107. (50) Fisichella, M.; Morini, L.; Sempio, C.; Groppi, A. Anal. Bioanal. Chem. 2014, 406, 3497−3506. (51) Laha, T. J.; Strathmann, F. G.; Wang, Z.; de Boer, I. H.; Thummel, K. E.; Hoofnagle, A. N. Clin. Chem. 2012, 58, 1711−1716. (52) Holtzapple, C. K.; Buckley, S. A.; Stanker, L. H. J. Agric. Food Chem. 1997, 45, 1984−1990. (53) Arevalo, J. H.; Taussig, M. J.; Wilson, I. A. Nature 1993, 365, 859−863.
F
DOI: 10.1021/acs.analchem.5b01895 Anal. Chem. XXXX, XXX, XXX−XXX