Anal. Chem. 2011, 83, 297–302
Quantitative Detection of Human Tumor Necrosis Factor r by a Resonance Raman Enzyme-Linked Immunosorbent Assay Stacey Laing,† Aaron Hernandez-Santana,† Jo¨rg Sassmannshausen,‡ Darren L. Asquith,§ Iain B. McInnes,§ Karen Faulds,*,† and Duncan Graham*,† Centre for Molecular Nanometrology, WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street, Glasgow G1 1XL, and Division of Infection, Immunity and Inflammation, Glasgow Biomedical Research Centre, University of Glasgow, 120 University Place, Glasgow G12 8TA, United Kingdom Tumor necrosis factor r is an inflammatory cytokine which has been linked with many infectious and inflammatory diseases. Detection and quantification of this key biomarker is commonly achieved by use of an enzymelinked immunosorbent assay (ELISA). This fundamental technique uses the spectroscopic detection of a chromogen such as 3,3′,5,5′-tetramethylbenzidine (TMB). Horseradish peroxidase (HRP), bound to the detection antibody, catalyzes the oxidation of TMB by hydrogen peroxide to generate colored products which may be measured spectrophotometrically. In this study we have used a conventional ELISA kit and shown that, by replacing the traditional colorimetric detection with resonance Raman spectroscopy, we can achieve 50 times lower detection limits and the potential for multiplexed analysis is increased. In this approach, the laser wavelength was tuned to be in resonance with an electronic transition of the oxidized TMB. The relative intensity of the enhanced Raman bands is proportional to the amount of TMB, thus providing a means of improved quantification. Furthermore, TMB is one of the most widely used chromogenic substrates for HRP-based detection and commercial ELISA test kits, indicating that this detection technique is applicable to a large number of target analytes. Tumor necrosis factor R (TNF-R) is an inflammatory cytokine produced by cells of the immune system in response to infection or cancer. When overproduced, it plays a major role in chronic inflammatory diseases such as rheumatoid arthritis,1 psoriasis,2 and Crohn’s disease.3 It has also been linked with conditions such * To whom correspondence should be addressed. Fax: +44 (0)141 548 4787. Phone: +44 (0)141 548 4701. E-mail:
[email protected]. † Centre for Molecular Nanometrology, WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde. ‡ WestCHEM, Department of Pure and Applied Chemistry, University of Strathclyde. § Division of Infection, Immunity and Inflammation, Glasgow Biomedical Research Centre, University of Glasgow. (1) Tetta, C.; Camussi, G.; Modena, V.; Di Vittorio, C.; Baglioni, C. Ann. Rheum. Dis. 1990, 49, 665–667. (2) Ettehadi, P.; Greaves, M. W.; Wallach, D.; Aderka, D.; Camp, R. D. R. Clin. Exp. Immunol. 1994, 96, 146–151. (3) D’Haens, G. Curr. Pharm. Des. 2003, 9, 289–294. 10.1021/ac1024039 2011 American Chemical Society Published on Web 12/01/2010
as septic shock syndrome,4 diabetes,5 and preeclampsia.6 A recent study involving healthy individuals found that 50 of the 58 samples analyzed were below the detection limit (2 pg/mL) of a conventional enzyme-linked immunosorbent assay (ELISA).7 Urszula et al.6 reported increased levels of TNF-R in serum from 0.7 ± 0.3 pg/mL in healthy pregnant women to 1.5 ± 0.7 pg/mL in women with preeclampsia. These levels of the cytokine are also below the detection limit of conventional tests, and thus, a method which offers a lower detection limit is advantageous. ELISA is the most common clinical method for the detection and quantification of important biomarkers.8 This fundamental technique uses antibody-antigen specific binding where an antibody immobilized on a solid substrate will react with its corresponding antigen, which will then react with an enzymelabeled antibody. The enzyme activity can consequently be measured by means of its reaction with a chromogenic substrate to generate a measurable signal which can be correlated to the amount of biomarker present in a sample. Common enzyme labels are horseradish peroxidase (HRP), alkaline phosphatase (ALP), and β-galactosidase.9 HRP is the smallest and most stable of the three and therefore the most desirable.10 The faster catalytic rate of HRP also makes it more sensitive than ALP as it generates a strong signal in a short time. However, this means that HRP reactions can be self-limiting due to substrate inhibition of the enzyme.8 HRP catalyzes the oxidation of many reducing substrates by H2O2. The most widely used substrate in HRP detection systems is 3,3′,5,5′-tetramethylbenzidine (TMB). TMB is less toxic and more sensitive than alternative substrates such as O-phenylenediamine (OPD) and 2,2′-azino-bis(3-ethylbenzthia(4) Mira, J. P.; Cariou, A.; Grall, F.; Delclaux, C.; Losser, M. R.; Heshmati, F.; Cheval, C.; Monchi, M.; Teboul, J. L.; Riche´, F.; Leleu, G.; Arbibe, L.; Mignon, A.; Delpech, M.; Dhainaut, J. F. J. Am. Med. Assoc. 1999, 282, 561–568. (5) Galic, S.; Oakhill, J. S.; Steinberg, G. R. Mol. Cell. Endocrinol. 2010, 316, 129–139. (6) Urszula, T. M.; Jerzy, L.; Bozena, K.; Ewa, F.; Agata, K.; Anna, N.; Tomasz, M. Ginekol. Pol. 2010, 81, 192–196. (7) Thermo Scientific Human TNF-a ELISA Kit (EH3TNFA) Product Instruction Manual; Pierce Biotechnology: Rockford, IL, 2009. (8) Porstmann, T.; Kiessig, S. T. J. Immunol. Methods 1992, 150, 5–21. (9) Wisdom, G. B. Clin. Chem. 1976, 22 (8), 1243–1255. (10) Deshpande, S. S. Enzyme Immunoassays: From Concepts to Product Development; Kruwer Academic Publishers: Dordrecht, The Netherlands, 1996.
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
297
Scheme 1. Oxidation of TMB by H2O2 in the Presence of HRP Proceeds by Two Successive One-Electron Oxidation Stepsa
a The radical intermediate TMB+ coexists in solution with a blue CTC, attributed to the reversible complexation of TMB0 and TMB2+.
zoline-6-sulfonic acid) (ABTS), and thus, its use as a substrate in HRP-based detection systems has been studied extensively. The HRP-catalyzed oxidation of TMB0 by H2O2 proceeds via the two-step two-electron reaction shown in Scheme 1. The first step (one-electron oxidation) yields a radical cation which exists in rapid equilibrium with a blue charge transfer complex (CTC). Formation of this species is attributed to the reversible complexation of the parent diamine (TMB0, electron donor) and the diimine (TMB2+, electron acceptor). At low pH the formation of the two-electron product is favorable; therefore, addition of a strong inorganic acid such as H2SO4 can terminate the reaction, yielding the yellow TMB2+. Quantitative analysis of the colored products may be carried out on a spectrophotometer, thus providing a convenient means of detection. To achieve lower detection limits, a wider dynamic range, and shorter incubation times, alternative detection methods have been investigated to replace the traditional colorimetric detection. As well as its chromogenic properties, TMB is also electroactive, and thus, its use as an electrochemical substrate has also been investigated. Volpe et al. utilized TMB as an electrochemical substrate for direct detection of HRP and achieved a lower detection limit of 8.5 × 10-14 M.11 They compared three substrates, TMB, hydroquinone, and p-aminophenyl phosphate (PAPP), and concluded that TMB was the best for low-level detection of HRP. Fanjul-Balado et al. also made use of its electrochemical properties and applied amperometric detection to a conventional ELISA system for the detection of HRP at levels as low as 2 × 10-14 M.12 They also used the system indirectly for the quantification of the protein interleukin-6 (IL6) in the range from 3.12 to 300 pg/mL.12 Electrochemical (11) Volpe, G.; Compagnone, D.; Draisci, R.; Palleschi, G. Analyst 1998, 123, 1303–1307. (12) Fanjul-Balado, P.; Gonza´lez-Garcı´a, M. B.; Costa-Garcı´a, A. Anal. Bioanal. Chem. 2005, 382, 297–302.
298
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
monitoring of the system was further utilized by Baldrich et al. in their development of an amperometric biosensor for detection of HRP/H2O2 to levels as low as 6 fM HRP or 5.4 µM H2O2.13 Although the electrochemical methods give good detection limits, simple instrumentation, and low production costs, they also show limitations such as variability between arrays, random surface coverage, and the inconvenience of having to activate the electrode between steps.12,13 Furthermore, the presence of large-sized anions commonly found in patient test samples such as heparin and DNA have been shown to affect the electro-oxidation of TMB.14 Raman spectroscopy is a selective and noninvasive method of chemical analysis that enables real-time reaction monitoring and characterization of compounds.15 This spectroscopic method gives unique vibrational fingerprints in the form of sharp, well-resolved bands which can be used to distinguish between structurally similar molecules. However, one of the main drawbacks of this technique is the lack of sensitivity. The selectivity and sensitivity of Raman scattering may be improved by tuning the excitation frequency close to, or into resonance with, an electronic excited state of the molecule. This effect is known as resonance Raman scattering (RRS), and enhancements of certain Raman bands of up to 106 orders of magnitude have been reported.16 Furthermore, only Raman bands originating from the chromophore of the analyte are enhanced, which can afford selective detection of a resonant analyte in the presence of nonresonant contaminants. The main drawback and limitation is the possibility of interference arising from a fluorescing analyte, which can often swamp the Raman signal. In this study we show how RRS can be used as the detection technique in a TMB/HRP ELISA system, with significant advantages over standard TMB detection. To demonstrate this, a conventional ELISA kit was used and a comparison between the traditional colorimetric detection method and RRS for the detection and quantification of TNF-R performed. Colorimetric detection was selected for comparison as this allowed a direct assessment, without alteration of any reagents, of how RRS could be used as a new method of detection and quantification of a common substrate system. The system was successfully applied to control samples, as well as patient blood serum samples, indicating that the method is applicable in a clinical scenario. In addition, we report the first resonance Raman spectra of the charge transfer complex. EXPERIMENTAL SECTION Chemicals and Instrumentation. An ELISA test kit and control solutions were purchased from R&D Systems Inc. (United Kingdom). The ELISA kit comprised a microtiter plate coated with anti-TNF-R (monoclonal), anti-TNF-R-HRP conjugate (polyclonal), hydrogen peroxide, TMB, 2 N sulfuric acid, a surfactant wash buffer (WB), and a lyophilized TNF-R standard. Hydrogen peroxide and TMB were mixed before use to yield the substrate (13) Baldrich, E.; del Campo, F. J.; Munoz, F. X. Biosens. Bioelectron. 2009, 25, 920–926. (14) Liu, M. L.; Zhang, Y.; Chen, Y.; Xie, Q.; Yao, S. J. Electroanal. Chem. 2008, 622, 184–192. (15) Smith, E., Dent, G., Eds. Modern Raman Spectroscopy: A Practical Approach, John Wiley & Sons, Chichester, U.K., 2005; 210 pp. (16) Jensen, L.; Zhao, L. L.; Autschbach, J.; Schatz, G. C. J. Chem. Phys. 2005, 123, 174110.
solution. TNF-R standard (10 ng) was reconstituted with doubly distilled deionized water (1 mL) to yield a 10 ng/mL stock solution. Dilutions of this standard were made up in animal serum (Calibrator Diluent RD6-35), also provided in the kit. Control samples (QC01-1, R&D Systems Inc.) contained a mixture of 13 recombinant human cytokines at low, medium, and high levels in porcine serum. The controls were also supplied lyophilized and reconstituted with doubly distilled deionized water. Patient samples were obtained from Darren Lee Asquith, University of Glasgow. The serum samples had already been processed as required and were therefore ready for analysis. Absorbance measurements were recorded using a microplate reader (Labsystems IEMS Reader MF) at 620 nm for the oneelectron oxidation products and at 450 nm for the two-electron oxidation product. UV-vis absorption spectra were acquired using a Cary 300 Bio UV-vis spectrophotometer. Measurements were carried out in a disposable plastic cuvette containing the ELISA product of the TNF-R standard (200 µL, 250 pg/mL) and doubly distilled deionized water (300 µL). Samples were scanned from 300 to 800 nm. Raman spectra were collected using a Renishaw inVia Raman microscope with an argon ion laser (514 nm) and a Renishaw Ramascope with a HeNe laser (633 nm). Samples were analyzed in microtiter plates using a 20× (NA 0.5) long-working distance objective. WiRE 2.0 software (Renishaw PLC) was used to run both Raman spectrometers. Each sample was prepared in triplicate, and three spectra (3 × 3 s) were recorded for each. Spectra were baseline corrected using Grams software (AI 7.00), and the average peak intensities were calculated. Microsoft Excel was used for the manipulation of the data. Computational studies were conducted using the program GAMESS, version Jan. 2009 (R1),17 or Gaussian03, revision E.01.18 Geometries have been fully optimized without symmetry constraints, involving the functional combinations according to Becke (hybrid) and Lee, Yang, and Parr (denoted B3LYP)19 as implemented in the program, or in the case of the charge transfer complex with the BP86 functional19,20 in connection with Grimme’s dispersion correction21 (denoted BP86-D). For all calculations the Pople standard 6-311G(d) basis set was used for all atoms. All obtained stationary points were subject to frequency calculation to confirm the minimum on the EPS (no imaginary frequencies). (17) Schmidt, M. W.; Baldridge, K. K.; Boatz, J. A.; Elbert, S. T.; Gordon, M. S.; Jensen, J. H.; Koseki, S.; Matsunaga, N.; Nguyen, K. A.; Su, S. J.; Windus, T. L.; Dupuis, M.; Montgomery, J. A. J. Comput. Chem. 1993, 14 (11), 1347–1363. (18) Frisch, M. J. ; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery., J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.; Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.; Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul, A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; Pople, J. A. Gaussian03, revision E.01; Gaussian, Inc.: Pittsburgh, PA, 2004. (19) Becke, A. D. Phys. Rev. A 1988, 38 (6), 3098–3100. (20) Perdew, J. P. Phys. Rev. B 1986, 33 (12), 8822–8824. (21) Grimme, S. J. Comput. Chem. 2006, 27 (15), 1787–1799.
Raman spectra were calculated with Gaussian at the B3LYP/6311G(d) level of theory. Enzyme-Linked Immunosorbent Assay. The ELISA test kit described above was used in accordance with the manufacturer’s instructions (R&D Systems Inc.). TNF-R standards, control samples, and human blood serum samples (200 µL) were incubated for 2 h at room temperature in the wells containing immobilized mouse monoclonal antibody against TNF-R. After incubation, the wells were aspirated and washed four times with WB. Anti-TNF-R-HRP conjugate (200 µL) was added to each well and incubated for 2 h at room temperature. Following another series of wash steps with WB, substrate solution (TMB/H2O2, 200 µL) was added, and the reaction was left to proceed for 20 min at room temperature. Optical density and Raman measurements of the blue CTC were carried out at this stage. For the yellow diimine oxidation product (TMB2+), analysis took place after the addition of 2 N sulfuric acid solution (50 µL). RESULTS AND DISCUSSION Early studies by Josephy et al. used optical and spin resonance spectroscopies to study TMB oxidation by a range of peroxidases.22 They achieved an electron spin resonance signal corresponding to the radical cation and a shift in optical spectra upon dilution, indicative of the CTC. Marquez and Dunford used stopped flow techniques to obtain rapid scan optical spectra and transient and steady-state kinetic data to show for the first time that TMB oxidation by peroxidase proceeds by two successive one-electron oxidations.23 TMB oxidation by HRP in the presence of H2O2 is characterized by the generation of colored products. We first monitored these changes using absorption spectroscopy (Figure 1a). During the first one-electron oxidation process, two absorption bands appear at around 370 and 650 nm, giving the solution a distinct blue color. The absorption band at 650 nm may be attributed exclusively to the CTC, with an extinction coefficient of 3.9 × 104 M-1 cm-1.22 Further oxidation leads to the formation of the diimine (yellow) and the replacement of these bands with an adsorption band at 450 nm (ε450 ) 5.9 × 104 M-1 cm-1). A green solution is sometimes observed, which results from a mixture of the initial blue form and the final yellow product. Resonance Raman Study of TMB Oxidation. Raman spectroscopy was used to study spectral changes using two common visible laser excitation wavelengths, 514 and 633 nm. Spectra were obtained using different laser wavelengths to investigate the possibility of selective resonance enhancement of particular chemical species. Excitation at 514 nm was selected as this wavelength is nonresonant with both the one-electron and twoelectron oxidation products. However, excitation at 633 nm of the one-electron oxidation products is in resonance with the lower electronic energy transition, which again is assigned exclusively to the CTC. Parts b and c of Figure 1 show the Raman spectra of the one-electron and two-electron oxidation products, respectively. The Raman bands at 1103, 1413, and 1436 cm-1 were constant throughout all the different spectra, which we assign to Raman bands arising from an excess of the parent, unoxidized diamine, (22) Josephy, P. D.; Eling, T.; Mason, R. P. J. Biol. Chem. 1982, 257, 3669– 3675. (23) Marquez, L. A.; Dunford, H. B. Biochemistry 1997, 36, 9349–9355.
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
299
Figure 1. (a) Absorption spectra of one-electron (solid line) and two-electron (dashed line) oxidation products of TMB0 by HRP in the presence of H2O2. (b, c) Raman spectra of the one-electron oxidation product (TMB+ + CTC) and of the two-electron oxidation product (TMB2+) at 514 and 633 nm. The resonance-enhanced Raman bands are marked with asterisks (b). The spectra have been normalized and shifted for illustration.
in solution. However, by tuning the laser wavelength to be in resonance with the lower energy transition (CTC) of the oneelectron oxidation products, the relative intensity of the Raman bands at 1191, 1336, and 1609 cm-1 was enhanced. This enhancement was not observed when using a nonresonant excitation wavelength (514 nm). To our knowledge, the RRS spectrum of the lower energy transition (CTC) has not been reported. However, RRS of the electrochemical oxidation processes of TMB in organic solvents has been reported previously by Misono et al.24 Unfortunately, these were only probed using 488 nm excitation, which was not in resonance with the CTC but with the TMB2+. They achieved spectra similar to those obtained in this work with shifts in frequency which can be attributed to the electron rearrangement between the one-electron and two-electron oxidation products. Resonance Raman studies of monocation radicals of similar compounds such as benzidine (BZ) and N,N,N′,N′-TMB (N-TMB) have also been reported previously.25,26 Computational studies were carried out to calculate the theoretical Raman frequencies of the CTC species. The calculated frequencies were in good agreement with the bands observed experimentally, and these were used to tentatively assign the bands as the following: CH3 bending modes (1191 cm-1), inter-ring C-C stretching modes (1336 cm-1), and combination of ring stretching and CH bending modes (1609 cm-1) (Table 1). Quantification of TNF-r Using RRS. The resonanceenhanced intensity of these three bands was proportional to the increasing concentration of TNF-R, which allowed for the quantification of the cytokine. The ELISA was carried out on a dilution series of the TNF-R standard (0-200 pg/mL), and the resonance (24) Misono, Y.; Ohkata, Y.; Morikawa, T.; Itoh, K. J. Electroanal. Chem. 1997, 436, 203–212. (25) Boilet, L.; Buntinx, G.; Lapouge, C.; Lefemeux, C.; Poizat, O. Phys. Chem. Chem. Phys. 2003, 5, 834–842. (26) Guichard, V.; Bourkba, A.; Poizat, O.; Buntinx, G. J. Phys. Chem. 1989, 93, 4429–4435.
300
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
Table 1. Tentative Assignments of the Resonance-Enhanced Raman Bands for the CTC Complex at 633 nm wavenumber (cm-1)
band assignment
1191 1336 1609
CH3 bending modes inter-ring C-C stretching modes ring stretching and C-H bending modes
Raman spectra of the CTC were recorded at 633 nm. By plotting the intensity of the enhanced Raman bands against the concentration of TNF-R on a log scale, a direct relationship could be obtained. The plot gave a good correlation coefficient (R2 ) 0.9984) and small error margins. This could then be used to calculate the lower detection limit of the assay, which was equal to the mean value of the blank plus 2 times the standard deviation. Figure 2a shows the plot of peak intensity against TNF-R concentration for the enhanced Raman band at 1191 cm-1. This wavenumber was selected as it gave the lowest detection limit of the three enhanced Raman bands. Absorbance measurements of the ELISA products were also taken in parallel at 620 nm in a spectrophotometer for the CTC, and the results were processed in the same way (Figure 2b). The detection limit obtained using the resonance Raman method was 90 fg/mL TNF-R, which was a 50-fold improvement compared to that of the conventional method, 4.50 pg/mL. The intensity of the Raman band at 1191 cm-1 for the TNF-R standards at the lower end of the concentration range can be observed in Figure 2b. We observed a linear increase in peak intensity with increasing concentration of TNF-R with the exception of the 0.049 pg/mL TNF-R standard. This indicates that the lower detection limit of the assay lies between 0.049 and 0.195 pg/mL, which is in good agreement with the calculated limit. Similarly, the optical density values obtained using the absorbance method for the low end of the concentration range can be found in Figure 2d. This chart shows that, using the absorbance method, it is not possible to
Figure 2. (a) Peak intensity against TNF-R concentration on a log scale for the enhanced Raman band at 1191 cm-1, recorded at 633 nm. (b) Chart of the intensity of the same peak for the TNF-R standards at the lower end of the concentration range. (c, d) Equivalent charts for the absorbance measurements, recorded at 620 nm.
distinguish between the standards of low concentration and the blank control and that the detection limit lies between 3.125 and 12.5 pg/mL. Again, this is in good agreement with the calculated limit of detection. To test the dynamic range of the assay, the dilution series was extended (0-4000 pg/mL) and the same method was followed. The range obtained using our system was slightly wider than that of the absorbance method. This is because, at absorbances greater than ∼1.2, the linear relationship between concentration and absorbance is lost. RRS was able to detect higher levels without similar topping out of the detection system. However, as a result of the self-limiting nature of the HRP in this reaction, neither method could successfully quantify levels as high as 4000 pg/ mL. This is not considered an issue in this work as the aim is to detect low levels for disease detection. Also, dilution of concentrated samples is always possible if necessary. The amount of TNF-R in controls containing a mixture of 13 recombinant human cytokines at low, medium, and high levels was then measured using both colorimetric detection and RRS. This allowed for a direct comparison of the two methods for detecting and quantifying TNF-R in the presence of serum proteins and other cytokines. The ELISA was carried out using a series of standards (15.6-1000 pg/mL), and the Raman spectra (CTC at 633 nm) and absorbance measurements (TMB2+ at 450 nm and CTC at 620 nm) were recorded. Results for each method were used to plot a standard curve against which the concentrations of the control samples, treated as unknowns, could be calculated. Table 2 shows the TNF-R concentrations obtained via each method with the expected range for each control. The expected values were taken from the human TNF-R ELISA kit product information booklet, as calculated by the manufacturer. All of the values obtained were within the expected value range. The presence of serum proteins and other cytokines failed to affect the quantification ability of the RRS method, thus proving its selectivity.
Table 2. TNF-r Concentrations Present in Low, Medium, and High Controls Obtained Using Both RRS and Colorimetric Detection Methods colorimetric colorimetric resonance Raman expected CTC, 620 nm TMB2+, 450 nm CTC, 633 nm value (±3 SD) control (pg/mL) (pg/mL) (pg/mL) (pg/mL) low 105.17 ± 5.20 medium 330.17 ± 18.93 high 669.33 ± 30.14
100.10 ± 3.77 312.81 ± 17.33 638.85 ± 24.72
111.41 ± 6.84 408.94 ± 15.56 734.57 ± 93.22
88-154 235-411 484-762
Table 3. TNF-r Concentrations Present in Eight RA Patients patient
colorimetric detection (pg/mL)
resonance Raman spectroscopy (pg/mL)
1 2 3 4 5 6 7 8
13.27 ± 1.15 12.76 ± 1.35 13.52 ± 1.56 12.50 ± 0.66 12.88 ± 0.77 12.12 ± 0.77 10.70 ± 0.9 11.47 ± 1.97
13.34 ± 3.88 12.59 ± 2.47 18.42 ± 4.05 17.19 ± 4.16 18.06 ± 2.13 13.37 ± 0.58 14.67 ± 2.52 11.77 ± 1.91
Finally, and to assess the compatibility of our method in a clinical setting, the blood serum of eight patients diagnosed with rheumatoid arthritis (RA) was analyzed using both methods for the quantification of TNF-R. The patients had been administered disease-modifying antirheumatic drugs (DMARDs), and the levels of the cytokine present in their blood was unknown. In a method similar to that used to analyze the controls, the patient samples were treated as unknowns and the concentrations calculated against a standard curve, generated in the same experiment. The values obtained both for the colorimetric detection method and using resonance Raman spectroscopy are shown in Table 3. The colorimetric detection method used in this experiment is what would normally be used for the analysis of clinical samples. Since our method gave results comparable to those of the colorimetric Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
301
detection, it is proven that our method is also suitable for the quantification of TNF-R in human blood serum. Although there are a few discrepancies in the results, these can be attributed to the instrumental setup and the time delay between collecting Raman scattering measurements. In the colorimetric detection method each reading is taken simultaneously using a microplate reader, whereas in the RRS method, measurements were taken one after the other by focusing the microscope in each separate well and moving the plate manually between samples. The consequent time delay in this approach could also allow for further oxidation of the CTC complex into the nonresonant TMB2+ species. This can account for the larger standard deviation found in some cases when the conventional method is replaced with RRS. This does not, however, present any real problem, as similar plate readers are available for Raman instruments. Such equipment could be exploited in a clinical environment to provide further accuracy as well as to reduce the duration of analysis. CONCLUSIONS We propose the use of RRS as a feasible way of quantifying TMB in solution. Use of this technique does not require modification of the standard ELISA procedure while providing lower detection limits. The dynamic range is also good and comparable to that of the colorimetric detection method. This was demonstrated using a conventional ELISA kit for TNF-R but should be applicable to any TMB-based assays. Since TMB is a common
302
Analytical Chemistry, Vol. 83, No. 1, January 1, 2011
substrate, we envisage that the system will be applicable to a variety of other biomarkers as well as the TNF-R used in this work. We also present the first resonance Raman spectra of the CTC at 633 nm, which to our knowledge has not yet been reported. One of the many benefits of this selective technique is that only bands associated with the colored CTC are enhanced. Furthermore, resonance Raman spectra were not affected in the presence of serum or other cytokines and gave results comparable to those of methods currently utilized, making this method suitable in a clinical scenario. Unlike standard absorbance measurements, resonance Raman spectra feature sharp bands, with potential for multiplexed analysis. ACKNOWLEDGMENT We acknowledge the Analytical Chemistry Trust Fund and EPSRC for the award of an analytical studentship to S.L. SUPPORTING INFORMATION AVAILABLE Resonance Raman spectra of the ELISA products of a dilution series of TNF-R standards (0-200 pg/mL) at 633 nm laser excitation wavelength. This material is available free of charge via the Internet at http://pubs.acs.org.
Received for review November 13, 2010. AC1024039
September
9,
2010.
Accepted
