Detection of Total and A1c-Glycosylated Hemoglobin in Human Whole

Sep 11, 2012 - Whole Blood Using Sandwich Immunoassays on. Polydimethylsiloxane-Based ... blood sugar level than glucose measurement. Thus, the...
0 downloads 0 Views 2MB Size
Download PDF
Article pubs.acs.org/ac

Detection of Total and A1c-Glycosylated Hemoglobin in Human Whole Blood Using Sandwich Immunoassays on Polydimethylsiloxane-Based Antibody Microarrays Huang-Han Chen,† Chih-Hsing Wu,‡ Mei-Ling Tsai,§ Yi-Jing Huang,§ and Shu-Hui Chen*,†,⊥ †

Department of Chemistry, National Cheng Kung University, Tainan 701, Taiwan College of Medicine, National Cheng Kung University Hospital, Tainan 701, Taiwan § Institute of Physiology, Medical College, National Cheng Kung University, Tainan 701, Taiwan ⊥ Agricultural Biotechnology Center, National Chung Hsing University, Taichung, 40227, Taiwan ‡

ABSTRACT: The percentage of glycosylated hemoglobin A1c (%GHbA1c) in human whole blood indicates the average plasma glucose concentration over a prolonged period of time and is used to diagnose diabetes. However, detecting GHbA1c in the whole blood using immunoassays has limited detection sensitivity due to its low percentage in total hemoglobin (tHb) and interference from various glycan moieties in the sample. We have developed a sandwich immunoassay using an antibody microarray on a polydimethylsiloxane (PDMS) substrate modified with fluorinated compounds to detect tHb and glycosylated hemoglobin A1c (GHbA1c) in human whole blood without sample pretreatment. A polyclonal antibody against hemoglobin (Hb) immobilized on PDMS is used as a common capture probe to enrich all forms of Hb followed by detection via monoclonal anti-Hb and specific monoclonal anti-GHbA1c antibodies for tHb and GHbA1c detection, respectively. This method prevents the use of glycan binding molecules and dramatically reduces the background interference, yielding a detection limit of 3.58 ng/mL for tHb and 0.20 ng/mL for GHbA1c. The fluorinated modification on PDMS is superior to the glass substrate and eliminates the need for the blocking step which is required in commercial enzyme linked immunosorbent assay (ELISA) kits. Moreover, the detection sensitivity for GHbA1c is 4− 5 orders of magnitude higher, but the required sample amount is 25 times less than the commercial method. On the basis of patient sample data, a good linear correlation between %GHbA1c values determined by our method and the certified high performance liquid chromatography (HPLC) standard method is shown with R2 > 0.98, indicating the great promise of the developed method for clinical applications.

G

spectroscopy (SERS),18 and drop coating deposition Raman spectroscopy (DCDRS),19 have been developed for detecting GHbA1c. However, only a few of these methods can be applied for clinical diagnostics. The concentration of total hemoglobin (tHb) in the whole blood is extremely high. If tHb is not separated from undepleted blood, the signal of the low abundant GHbA1c will likely be blocked by the strong signal of hemoglobin (Hb). Moreover, glycan moieties are ubiquitously present in biological fluids, resulting in interference for any glycan-based detection method. To date, ionexchange chromatography has been recognized as the gold standard for the determination of %GHbA1c. Compared to separation-based methods, nonseparation methods are more promising for use in point-of-care diagnostics. Immunoassays using a single specific GHbA1c antibody coupled with a field effect transistor20 and electrochemistry using a boric acid coated electrode21 have been used to detect GHbA1c. Whereas the former method has not been demonstrated with complicated biological samples, the latter is sensitive but

lycosylated hemoglobin A1c (GHbA1c) is generated from the reaction between glucose and the N-terminal valine of one or both hemoglobin beta-chains. The resulting product, aldimine, rearranges to form a stable ketoamine.1 The amount of GHbA1c formation is proportional to the mean blood glucose concentration to which the erythrocytes are exposed. Due to the limited lifetime of red blood cells (approximately 120 days), the percentage of GHbA1c in total hemoglobin (%GHbA1c) can indicate a mean value of blood glucose over approximately three months.2 This method is considered to be more accurate and objective for indicating the blood sugar level than glucose measurement. Thus, the measurement of %GHbA1c has become an important indicator for the diagnosis and treatment of diabetes.3−8 The clinical reference range for GHbA1c is 5−15%. The American Diabetes Association (ADA) has suggested using 6.5%GHbA1c as the cutoff value for a healthy individual.9 Many analytical techniques, including separation-based methods such as ion-exchange chromatography,10 boronate affinity chromatography,11 and isoelectric focusing12 and nonseparation-based methods such as mass spectroscopy,13 piezoelectric methods,14 electrochemical methods,15 colorimetry,16 immunoassays,17 surface enhanced resonance Raman © 2012 American Chemical Society

Received: June 24, 2012 Accepted: September 11, 2012 Published: September 11, 2012 8635

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641

Analytical Chemistry

Article

aminopropyltriethoxysilane (APTES) were obtained from Sigma (St. Louis, MO, USA). Acrylate-poly(ethylene glycol)N-hydroxysuccinimide (ACRL-PEG-NHS) (MW 5000) was obtained Laysan Bio (Arab, AL, USA). The photoinitiator 2,2dimethoxy-2-phenylacetophenone (DPA) was obtained from Fluka (Buchs, Switzerland). Phosphate buffered saline (PBS) was obtained from Pierce (Rockford, IL, USA). The commercial quality-control Level 1 and 2 materials were obtained from BIO-RAD (Irvine, CA, USA). The polyclonal goat antihemoglobin antibody was obtained from Novus (Littleton, CO, USA). The monoclonal mouse antihemoglobin and antihemoglobin A1c were obtained from Abcam (Cambridge, UK). The antimouse IgG-HRP was obtained from Santa Cruz (Santa Cruz, CA, USA). Acrylic acid (AA) was obtained from Fluka (Buchs, St. Gallen, Switzerland). The Enhanced Chemiluminescent Luminol Reagent-kit (ECL) was obtained from PerkinElmer Life Sciences (Boston, MA, USA) for detection, and the emission was captured by a digital imaging system (UVP Bio-Imaging Systems, CA, USA). PBST composed of 0.05% Tween 20 in PBS buffer was prepared in house. The human Hb and human GHbA1c ELISA kits were obtained from Bethyl (Montgomery, TX, USA) and Cusabio Biotech (Newark, DE, USA), respectively. Clinical Blood Sample Collection. Twenty-eight whole blood samples, 13 diabetic patients and 15 healthy subjects, were recruited from the out-patient clinic at National Cheng Kung University Hospital between 2011 and 2012. Moreover, all diabetic patients were examined by physicians to confirm their stable clinical condition. All collected blood samples were split into two fractions, and one fraction from each sample was sent to the certified lab for TOSOH G7 HPLC measurement. The other blood samples were kept at −80 °C until use. This study was approved by the Institute Review Board of National Cheng Kung University Hospital (IRB no: ER-100-033), and each subject signed the inform consent before examination. Human Hemoglobin and Hemoglobin A1c ELISA Kits. The commercial quality-control Level 1 and Level 2 standards were serially diluted by orders of magnitude ranging from 10−1 to 10−5, and the blood samples were diluted by an order of 10−3 with deionized water; the rest of the analysis followed the procedures provided by the commercial kits. Briefly, a volume of 100 μL of each diluted standard or blood sample was pipetted onto the commercial microplates, which were precoated with anti-GHbA1c or anti-Hb antibody. After incubation, biotin conjugated anti-GHbA1c was added to the microplate followed by HRP-avidin or anti-Hb followed by anti-IgG-HRP. Finally, TMB was added to each microwell, the solution was incubated in the dark at room temperature for 30 min, and then, the reaction was stopped by adding 100 μL of the stop solution to each microwell. The absorbance of each microwell was recorded by an ELISA reader (Tecan, CA, USA) at a wavelength of 450 nm. The blood samples (1 μL) were diluted with DI water to a volume of 1 mL, and 100 μL of the diluted blood sample was added to each microwell for analysis, following the same procedures as described for the standards. Fabrication of the PDMS-Based Antibody Microarray Chip. The fluorinated PDMS antibody microarray was fabricated following procedures described previously.25−27 Briefly, the PDMS oligomer was mixed with the curing agent at a weight ratio of 10:1 on a stainless steel template to form a 12 × 8 array with 2.5 mm id for each spot in a 5 cm × 4 cm space. Each spot on the surface was sequentially modified with 4 layers of branched PEI and PAA. The cross-linked PEI/PAA

suffers from interference arising from high blood sugar and albumin concentration, which can also be recognized by the borate coated on the electrode.21 Furthermore, the excellent accuracy and reproducibility achieved by proof-of-concept studies conducted with SERS and DCDRS opens substantive avenues for characterization and quantification of the glycosylation status of therapeutic proteins and holds potential for clinical applications in the future. However, to the best of our knowledge, the only clinical immunoassay using a single specific GHbA1c antibody is an assay based on the inhibition of latex agglutination22 that uses antibody-bound latex to agglutinate with the agglutinator, followed by inhibition of agglutination by the antigen present in the sample. Total hemoglobin is detected on the basis of an iron reduction mechanism, and the %GHbA1c value is reported as the output.22 Although separate sandwich enzyme linked immunosorbent assay (ELISA) assays for tHb and GHbA1c are commercially available from different companies, these kits are specified for research use only and we have not found any clinical report based on these products. Moreover, since tHb and GHbA1c are detected by separate products, it is not convenient to develop a standardized mean to convert the results to %GHbA1c value. In principle, a sandwich assay using two antibodies can greatly reduce nonspecific binding and is more sensitive than using a single antibody. However, two sitespecific epitopes are hard to map for a small glycosylation site. Alternatively, some sandwich assays use lectin or glycan-binding antibodies as the detection probe and the specific HbA1c antibody as the capture probe.23 However, using this method, blood sugar and glycosylated proteins can bind to the detection probe if not completely removed. The glycan group on the Fc domain of the immobilized capture antibody can also be recognized by lectins and cause high background interference unless chemical derivatization is used to block the glycan.24 In this study, we sought to develop a sandwich immunoassay for detecting both tHb and GHbA1c in human whole blood without sample pretreatment. Two novel parts of the overall process have been developed: (1) Instead of two specific antibodies or sugar binding molecules, we used a common capture probe to increase the recovery yield of the low percentage GHbA1c in the whole blood and used nonglycanbased molecules that recognize and differentiate tHb and GHbA1c as the detection probe. (2) For microarray fabrication, we will use the fluorinated compound-modified polydimethylsiloxane (PDMS) substrate that has been previously shown to resist nonspecific binding.25 Additionally, we compared the fluorinated modification on PDMS versus the glass substrate for tHb and GHbA1c detection. Furthermore, results obtained from the clinical samples were compared with the commercial GHbA1c ELISA kit and the standard method using TOSOH G7 high performance liquid chromatography (HPLC) to explore the clinical usefulness of the proposed platform.

■

EXPERIMENTAL SECTION Materials and Chemicals. The Sylgard 184 kit, containing vinyl-terminated PDMS base and curing agent, was acquired from Dow Corning (Midland, MI, USA). Hydrolyzed poly(styrene-alt-maleic anhydride) (h-PSMA) (MW 350 kDa), poly(ethyleneimine) (PEI) (MW 750 kDa), poly(acrylic acid) (PAA) (MW 100 kDa), 1-[3-(dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride (EDC), N-hydroxy-succinimide (NHS), tetramethylbenzidine (TMB), Tween 20, 1H,1H,2Hperfluoro-1-decene (FD), tetramethylbenzidine (TMB), and 8636

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641

Analytical Chemistry

Article

Figure 1. Workflow of the sandwich immunoassay for tHb and GHbA1c detection using the antibody microarray immobilized with the common capture probe, anti-Hb.

■

RESULTS AND DISCUSSION Sandwich Detection Scheme. As depicted in Figure 1, a common probe, polyclonal Hb antibody that can capture tHb including nonglycosylated and glycosylated Hb, is bound to the top of the modified PDMS substrate, and any unbound species, including many glycan-containing molecules, are washed away. The detection probes, i.e., the monoclonal antibodies against the Hb (anti-Hb) and GHbA1c motifs (anti-GHbA1c), are added to individual spots to bind to the captured Hb and GHbA1c, respectively. As shown, the monoclonal anti-Hb antibody recognizes an epitope (shown in red) that differs from the epitope (shown in green) recognized by the capture probe. Moreover, any serum glycoproteins remaining as nonspecific binders after the wash step or the glycan molecule on the Fc domain of the immobilized capture probe are not recognized by the anti-GHbA1c antibody (Figure 1). Subsequently, antimouse IgG-HRPs are added to individual spots to bind to the captured anti-Hb or anti-GHbA1c for chemiluminescence reaction and detection. We first tested the detection scheme with two commercial quality-control Level 1 (%GHbA1c = 5.6) and Level 2 (% GHbA1c = 9.3) standards. On the basis of the measurement using a commercial Hb ELISA kit (discussed below), the Level 1 standard was estimated to contain 358 μg/mL of tHb and 20.05 μg/mL of GHbA1c (5.6% of tHb). As shown in Figure 2A,B, the spot intensity for tHb and GHbA1c decreases as Standard Level 1 and Level 2 solution are serially diluted by an order of magnitude from 1/102, 1/103, and 1/104 to 1/105 (n = 3), indicating the proposed scheme can quantitatively detect and differentiate tHb and HbA1c proteins over a wide dynamic range up to 3 orders of magnitude. The detection limit was estimated to be approximately 3.58 ng/mL (S/N = 4.41) and 0.20 ng/mL (S/N = 3.41) for tHb and GHbA1c, respectively. The lower detection limit for GHbA1c is due to the lower background noise (enlarged region shown in Figure 2) detected

polyelectrolyte layers were photopolymerized with a mixture containing the FD (15% v/v), AA (1% v/v), and DPA photoinitiators (1% w/v). After being washed with anhydrous ethanol and dried under a nitrogen stream, the substrate was incubated with a protein G solution (20 μg/mL in PBS buffer) at room temperature for four hours. The protein G-coated PDMS microarray was immersed in 1 μg/mL of polyclonal goat antihemoglobin solution in PBST buffer for two hours and then washed with PBST to remove unbound species. Fabrication of the Glass-Based Antibody Chip. A volume of 80 μL of aminopropyltriethoxysilane (APTES) (10% v/v) in anhydrous ethanol was pipetted onto a bare glass substrate. Eighty microliters of ACRL-PEG-NHS (1000 μg/ mL) in PBS buffer (pH 7.4) solution was added to the activated glass surface followed by sequential layer-by-layer modifications with FD/AA, protein G, and polyclonal goat anti-Hb as described for PDMS. For glass substrate without fluorinated modification, the surface was modified as described but without the FD/AA layer. Immunoassays on Fluorinated PDMS/Glass Microarray. For immuno-analysis, 4 μL of the standard or diluted (10−3) whole blood solution was pipetted onto the PDMS antibody chip. The red blood cells (RBC) in the whole blood samples were not isolated, and RBCs were broken by DI water dilution without lysing techniques to obtain hemolysate solutions. After two hours of incubation, the monoclonal mouse anti-GHbA1c antibody or the mouse anti-Hb (0.5 μg/ mL in PBST buffer) was added to each spot followed by antimouse IgG-HRP (0.5 μg/mL in PBST buffer) and the ECL reagent or TMB reagent as that used in ELISA kits. The emitted chemiluminescence was detected by a BioSpectrum imaging system (UVP, Bio-Imaging Systems, CA, USA). The TMB absorbance was detected by the ELISA reader as described above. For the glass chip, 40 μL of the standard or diluted (10−3) whole blood solution was used for the analysis. 8637

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641

Analytical Chemistry

Article

Figure 2. Detection of tHb and GHbA1c proteins from serial dilutions (10−2, 10−3, 10−4, 10−5 ,and blank) of (A) Level 1 and (B) Level 2 standards with chemiluminescence detection and (C) the Level 1 standard (10−5 and blank) with TMB absorbance detection. Each spot is 2.5 mm in diameter, and the blank spot was spotted with the solvent only. Three repeated measurements (n = 3) were performed for each solution.

sensitivity acquired by TMB absorbance is comparable to that acquired by ECL emission. The ratios of emission intensity of HbA1c and tHb (IHbA1c/ItHb) detected from Level 1 (% GHbA1c = 5.6) and Level 2 (%GHbA1c = 9.3) standards were then used to establish a conversion equation %HbA1c (PDMS) = A*(100*IHbA1c/ItHb) + B to be used for the whole blood samples discussed in below. Fluorinated Modification on PDMS and Glass Substrate. Because the glass substrate is the most commonly used

from the blank spot compared to the background of tHb. As shown in Table 1, when the background noise was subtracted from the detected emission intensity, almost identical linear regression equations with R2 value >0.97 were obtained for tHb and HbA1c from the log (emission intensity) versus log (concentration) plots using data obtained from either Level 1 or Level 2 standards. TMB absorbance detection used for ELISA assay (discussed in below) was applied to the PDMS chips, and results shown in Figure 2C indicate that the 8638

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641

Analytical Chemistry

Article

higher surface coverage. These results demonstrate the usefulness of the fluorinated modification for immunoassays and its superior performance when coated on the PDMS substrate compared to the glass substrate. Comparison with Commercial GHbA1c ELISA Kits. This method was compared to the method using commercial Hb and GHbA1c ELISA kits. On the basis of the specifications, the Hb ELISA kit has a detection range from 10 to 100 ng/mL for tHb, which is comparable to our method. However, the specified detection range for the GHbA1c ELISA kit is 15.6− 1000 μg/mL which is poorer than our method by 5−6 orders of magnitude. Although the GHbA1c ELISA is specified to use two antibodies specific for GHbA1c as the capture and detection probe, there is no specification about the epitope of these two antibodies. As stated, it is hard to have two antibodies with different epitopes against a small glycan site (GA1c). In contrast, it is easy to have two specific antibodies against two different epitopes on the same protein such as Hb (red and green shown in Figure 1). We suspect the much poorer detection limit of the GHbA1c kit compared to the Hb kit may be due to the poorer specificity and sensitivity of the antibodies. The commercial Hb ELISA kit can be applied to detect Hb in the standards (see discussion above), but the GHbA1c ELISA kit can only barely detect GHbA1c in the Level 1 standard because the estimated GHbA1c concentration (20.05 μg/mL) in the Level 1 standard is very close to the low limitation of the specified detection range (15.6 μg/mL). In contrast, our fluorinated PDMS microarray could detect both tHb and GHbA1c in the diluted (10−5) Level 1 standard. We think the lower detection limit of GHbA1c obtained from our method is due to high recovery and less nonspecific binding. Our detection method has several advantages that lead to high sensitivity and a low detection limit for GHbA1c. First, compared to methods using monoclonal anti-GHbA1c as the single capture probe, our two-step method using polyclonal Hb antibody as the capture probe could greatly increase the recovery yield but without sacrificing the specificity of the

Table 1. Regression Equations of Level One and Level Two Standards Using Data Shown in Figure 2 log(intensity − blank intensity) = A + B log(dilution order) level one A B R2

level two

Hb

HbAlc

Hb

HbAlc

5.5325 −0.3493 0.9747

4.8821 −0.3285 0.9821

5.4756 −0.321 0.9711

5.2994 −0.3351 0.9719

substrate for microarray fabrication, we compared the detection sensitivity obtained from the fluorinated modification on the glass with the detection sensitivity obtained from the PDMS substrate. The glass substrate was activated by silanization followed by fluorinated coatings and covalent protein G binding through EDC/NHS chemistry using the layer-bylayer method for PDMS. As shown in Figure 3, without fluorinated modification, the glass surface can hardly detect any tHb or GHbA1c signal in the diluted (10−3) Level 1 standard. With the fluorinated modification, the background noise detected from the blank spot is reduced and the detection sensitivity of tHb is substantially increased (Figure 3), indicating the usefulness of the fluorinated modification in reducing nonspecific binding and background noise. Although the detection sensitivity for GHbA1c obtained from the glass substrate is significantly increased after the fluorinated modification, the detected signal is still substantially lower than that obtained from the modified PDMS substrate (Figure 3). Because the volume of the blood sample added to the glass substrate (40 μL) is 10 times more than that added to the PDMS substrate (4 μL), the detection sensitivity obtained from the modified glass substrate was estimated to be 300−400 times lower than that obtained from the modified PDMS chip. The cause of the higher detection sensitivity on the PDMS substrate compared to the glass substrate may be due to the formation of multicoated layers on the soft PDMS surface,22 and thus, a

Figure 3. Comparison of the detection sensitivity for the diluted (10−3) Level 1 standard using PDMS substrate modified with the fluorinated compounds as well as the glass substrate with and without fluorinated modification. The blank spot was spotted with the solvent without the standard. Three replicates (n = 3) were performed for each measurement. 8639

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641

Analytical Chemistry

Article

detection probe using GHbA1c. Second, compared to alternative methods using lectin or other glycan binding antibodies as the detection probe, any nonremoved glycocompounds that do not have an GHbA1c motif (marked as nonspecific binders in Figure 1) will not bind to the monoclonal anti-GHbA1c antibody and thus will not contribute to the background noise. Finally, the fluorinated modification on PDMS can resist nonspecific binding and thus leads to less nonspecific binding. In addition, the assay established on the modified PDMS microarray is much more reproducible (CV% ranging from 0.19 to 0.82%, n = 3) than that using the commercial GHbA1c ELISA kit (CV% ranging from 56.2 to 1.3%, n = 3). Furthermore, no blocking step is required for the modified PDMS chip, and the required sample volume is as little as 4 μL. In contrast, the blocking step is necessary for the commercial kit and more than 100 μL of the sample is required for the analysis. Whole Blood Analysis. A cohort of 28 clinical blood samples from 13 normal individuals and 15 diabetic patients was collected and analyzed by our method using the modified PDMS antibody microarray, the commercial ELISA kits, and the standard method using TOSOH G7 ion exchange HPLC for comparison. The %GHbA1c values determined by PDMS microarray were calculated from the conversion equation established by Level 1 and Level 2 standards (see above). The data show that the diabetic bloods can be discriminated from the healthy subjects. In the meantime, the %GHbA1c ELISA values were calculated from the ratios of the Hb and GHbA1c levels determined by the Hb and GHbA1c ELISA kits following the calibration equations specified by each kit, and the %GHbA1c was directly calculated from the concentrations determined by each kit. The %GHbA1c HPLC values of the same blood samples were provided by a certified lab. As shown in Figure 4A, the calculated %GHbA1c PDMS values correlated well with the values determined by the standard method using HPLC (%GHbA1c HPLC = 1.06*%GHbA1c PDMS − 0.46, R2 = 0.98). Moreover, as shown in Figure 4A, the CV values obtained from the blood samples (n = 3) are indistinguishable from the CV values obtained from the standards. In contrast, as shown in Figure 4B, no correlation was observed between the values obtained from the commercial ELISA kits and the HPLC method for 6 diabetic samples. We think the low sensitivity and specificity for GHbA1c in the whole blood is the major cause. First of all, these kits are specified for research use only. Second, the specified low end concentration of GHbA1c is close to its concentration in Level 1 standard or in the normal blood. In order to completely release GHbA1c, all patient bloods including those used in Figure 4B were diluted by deionized water (1/1000). Although these GHbA1c levels were still detectable (S/N > 3), they could be lower than the specified quantification range. Our method, however, could detect GHbA1c in the whole blood which is diluted up to 5 orders of magnitude (10−5). Furthermore, no blocking step was used by our method, but a blocking step was used for the ELISA kits as specified. Thus, we concluded that our method is superior to the commercial ELISA kits in sensitivity, reproducibility, and the ease of use. The good correlation of our results with the certified HPLC values (Figure 4A) also demonstrates that our sandwich immunoassays using antibody microarrays on the modified PDMS substrate have great potential for clinical applications. However, the current method requires 2 h of incubation or more and may impede application in clinics where immediate and high volume throughput is necessary. We

Figure 4. Correlation plots of the %GHbA1c values of the patient whole blood and Level 1 and 2 standards determined from (A) the modified PDMS antibody microarray versus the HPLC method and (B) the commercial ELISA kits versus the HPLC method. Three replicates (n = 3) were performed for each measurement with standard deviation indicated.

are currently working on reducing incubation time by improving the diffusion/mixing efficiency.

■

CONCLUSION By combining a detection scheme using a common capture probe and specific detection probes with antibody microarray on the fluoro-modified PDMS substrate, serum tHb and GHbA1c levels can be accurately and reproducibly determined without sample pretreatment and a tedious blocking step. Moreover, the determined %GHbA1c values can be correlated with the values determined by the standard HPLC method. In view of the low cost and biodegradable nature of PDMS soft material, we believe the reported method can be further developed for point-of-care diagnostics.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by the National Science Council in Taiwan (NSC 100-2321-B-006-008 and NSC 99-2113-M-006001-MY3) and in part by the Ministry of Education, Taiwan, ROC under the ATU plan, to Prof. S.-H.C. The authors thank Ms. Yu-Chen Shih and Ms. Pu-Hsian Guo for their administrative assistance. 8640

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641

Analytical Chemistry

■

Article

REFERENCES

(1) Park, J. Y.; Chang, B. Y.; Nam, H. H.; Park, S. M. Anal. Chem. 2008, 80, 8035−8044. (2) Goldstein, D. E.; Nathan, D.; Little, R. R.; Peterson, C. M.; Lorenz, R. A.; Sacks, D. B.; Malone, J. I. Diabetes Care 2004, 27, 1761−1773. (3) Sacks, D. B.; Bruns, D. E.; Goldstein, D. E.; Maclaren, N. K.; McDonald, J. M.; Parrott, M. Clin. Chem. 2002, 48, 436−472. (4) The Diabetes Control and Complications Trial Research Group. N. Engl. J. Med. 1993, 329, 977−986. (5) UK Prospective Diabetes Study (UKPDS) Group. Lancet 1998, 352, 837−853. (6) American Diabetes Association. Diabetes Care 2009, 32 (Suppl 1), S13−S62. (7) Petersen, J. R.; Omoruyi, F. O.; Mohammad, A. A.; Shea, T. J.; Okorodudu, A. O.; Ju, H. Clin. Chim. Acta 2010, 411, 2062−2066. (8) Hoelzel, W.; Miedema, K. J. Int. Fed. Clin. Chem. 1996, 9, 62−67. (9) Thevarajah, M.; Nadzimah, M. N.; Chew, Y. Y. Clin. Biochem. 2009, 42, 430−434. (10) Lafferty, J. D.; McFarlane, A. G.; Chui, D. H. K. Arch. Pathol. Lab. Med. 2002, 126, 1494−1500. (11) Frantzen, F.; Grimsrud, K.; Heggli, D. E.; Sundrehagen, E. J. Chromatogr., B 1995, 670, 37−45. (12) Mullins, R. E.; Austin, G. E. Clin. Chem. 1986, 32, 1460−1463. (13) Roberts, N. B.; Amara, A. B.; Morris, M.; Green, B. N. Clin. Chem. 2001, 47, 316−321. (14) Halámek, J.; Wollenberger, U.; Stöklein, W.; Scheller, F. M. Electrochim. Acta 2007, 53, 1127−1133. (15) Liu, S.; Wollenberger, U.; Katterle, M.; Scheller, F. W. Sens. Actuators, B 2006, 113, 623−629. (16) Adamczyk, M.; Chen, Y. Y.; Johnson, D. D.; Mattingly, P. G.; Moore, J. A.; Pan, Y.; Reddy, R. E. Bioorg. Med. Chem. Lett. 2006, 16, 1324−1328. (17) Stöllner, D.; Stöcklein, W.; Scheller, F.; Warsinke, A. Anal. Chim. Acta 2002, 470, 111−119. (18) Syamala Kiran, M.; Itoh, T.; Yoshida, K.; Kawashima, N.; Biju, V.; Ishikawa, M. Anal. Chem. 2010, 82, 1342−1348. (19) Barman, I.; Chari Dingari, N.; Woong, K. J.; Horowitz, G. L.; Dasari, R. R.; Feld, M. S. Anal. Chem. 2012, 84, 2474−2482. (20) Qu, L.; Bian, C.; Sun, J.; Han, J.; Xia, S. Proc. SPIE 2008, 7159, 7159081−7159087. (21) Song, Y. S.; Yoon, H. C. Sens. Actuators, B 2009, 1, 233−239. (22) Jean, S.; Nathalie, L.; Emmanuelle, L.; Philippe, G. Clin. Chem. Lab. Med. 2008, 46, 1195−1198. (23) Kuno, A.; Uchiyama, N.; Koseki-Kuno, S.; Ebe, Y.; Takashima, S.; Yamada, M.; Hirabayashi, J. Nat. Methods 2005, 2, 851−856. (24) Chen, S.; LaRoche, T.; Hamelinck, D.; Bergsma, D.; Brenner, D.; Simeone, D.; Haab, B. B. Nat. Methods 2007, 4, 437−444. (25) Chen, H. H.; Sung, W. C.; Liang, S. S.; Chen, S. H. Anal. Chem. 2010, 82, 7804−7813. (26) Makamba, H.; Hsieh, Y. Y.; Sung, W. C.; Chen, S. H. Anal. Chem. 2005, 77, 3971−3978. (27) Sung, W. C.; Chen, H. H.; Makamba, H.; Chen, S. H. Anal. Chem. 2009, 81, 7967−7973.

8641

dx.doi.org/10.1021/ac301756d | Anal. Chem. 2012, 84, 8635−8641