Anal. Chem. 2001, 73, 844-847
Technical Notes
Needle-Type Dual Microsensor for the Simultaneous Monitoring of Glucose and Insulin Joseph Wang* and Xueji Zhang
Department of Chemistry and Biochemistry, New Mexico State University, Las Cruces, New Mexico 88003
A miniature needle-type sensor suitable for the simultaneous amperometric monitoring of glucose and insulin is described. The integrated microsensor consists of dual (biologically and chemically) modified carbon-paste working electrodes inserted into a 14-guage needle. The glucose probe is based on the biocatalytic action of glucose oxidase, and the insulin one relies on the electrocatalytic activity of ruthenium oxide. The analytical performance of the dual sensor is assessed under flow injection conditions. The needle dual detector exhibits a very rapid response to dynamic changes in the concentrations of glucose and insulin. No apparent cross reactivity is observed in mixtures containing millimolar glucose levels and nanomolar insulin concentrations. The response is highly linear (to at least 1000 nM insulin and 14 mM glucose) and reproducible (RSD ) 2.6-4.1%). The combination microsensor holds great promise for real-time measurements of the insulin/glucose ratio and for improved management of diabetes. Diabetes mellitus is a chronic metabolic disorder that results from a total or partial deficiency of insulin. Much attention has, thus ,been given to the development of implantable electrochemical biosensors for the continuous monitoring of glucose.1-3 Such miniaturized devices are commonly based on the enzymatic oxidation of glucose by the immobilized enzyme coupled to amperometric detection of the liberated hydrogen peroxide. Tighter diabetes control would benefit from improved glucose monitoring and the simultaneous detection of glucose and other analytes (particularly insulin). This note describes the design and analytical performance of a needle-type probe, integrating an amperometric glucose biosensor with an electrocatalytic insulin microsensor. Multipleanalyte electrochemical sensors were reported previously for the simultaneous monitoring of potassium and carbon dioxide4 and glucose and lactate.5 Insulin sensing is of great importance for clinical diagnostics, because it serves as a predictor of diabetes * To whom correspondence should be addressed. E-mail:
[email protected]. (1) Reach, G.; Wilson, G. S. Anal. Chem. 1992, 64, 381A. (2) Henry, C. Anal. Chem. 1998, 70, 594A. (3) Wang, J. Anal. Chem. 1999, 71, 328R. (4) Collinson, M. E.; Aebli, J. P.; Petty, J.; Meyerhoff, M. E. Anal. Chem. 1989, 61, 2365. (5) Yang, Q.; Atanasov, P.; Wilkins, E. Electroanalysis 1998, 10, 752.
844 Analytical Chemistry, Vol. 73, No. 4, February 15, 2001
(glucose tolerance), of insulinoma, and trauma. 6-8 The simultaneous monitoring of insulin and glucose, and hence, of the insulin/ glucose ratio, is thus of great relevance for the diagnosis of insulinoma7 and the management of diabetes. Cox and Gray9 developed an insulin-sensitive chemically modified electrode, based on ruthenium-oxide/cyanoruthenate (RuO/CN-Ru) coating, that promotes the oxidation of insulin in acidic media. More attractive for insulin detection at physiological pH is the rutheniumoxide (RuOx)-type catalytic film developed by Gorski et al.10 The utility of the RuOx sensor for monitoring insulin secretion from pancreatic β cells was demonstrated.11 In the present study, we integrated, on the same needle body, an insulin-sensitive RuOxmodified carbon-paste microelectrode with a metalized (rhodinized) carbon amperometric glucose biosensor (Figure 1). The attractive performance characteristics of rhodium-dispersed carbonpaste oxidase biosensors were described earlier.12 Despite the substantial concentration difference [mM(glucose) and nM(insulin)] and the use of different transduction principles, the microsensors of the new combination probe respond independently and rapidly to the corresponding target analytes, with no apparent cross reactivity. The dual microsensor, thus, offers great promise for monitoring the time dependent variation of glucose and insulin and, hence, for simultaneously measuring in real time the insulin/glucose ratio. These and other attractive performance characteristics of the needle-type glucose/insulin microsensor are reported in the following sections. EXPERIMENTAL SECTION Apparatus. Cyclic voltammetry was carried out using an Autolab modular electrochemical system (Eco Chemie, Ultrecht, The Netherlands) equipped with a PSTA 10 module and driven by GPES software (Eco Chemie). Flow injection amperometric data were collected using two EG&G model 264A voltammetric analyzers in connection to two OmniScribe D500 strip-chart (6) Jenssen, T. G.; Toft, I.; Bonaa, K. H. Diabetologia 1995, 38, 127A. (7) Ravel, R. “Clinical Laboratory Medicine”, 5th edition; Mosby Year Book: St. Louis, 1989; p 661. (8) Tannuri, U.; Coelho, M. C.; Maksoud, J. G. Pediatr. Surg. Int. 1993, 8, 210. (9) Cox, J. A.; Gray, T. Anal. Chem. 1989, 61, 2462. (10) Gorski, W.; Aspinwall C. A.; Lakey J. R. T.; Kennedy R. T. J. Electroanal. Chem. 1997, 425, 191. (11) Kennedy, R. T.; Huang, L.; Aspinwall, C. A. J. Am. Chem. Soc. 1996, 118, 1195. (12) Wang, J.; Liu, J.; Chen, L.; Lu, F. Anal. Chem. 1994, 66, 3600. 10.1021/ac0009393 CCC: $20.00
© 2001 American Chemical Society Published on Web 01/19/2001
Figure 2. Alternative flow injections of samples containing 5 mM glucose (A) and 400 nM insulin (B). Applied potentials (for both sensors), +0.60 V vs stainless steel (glucose) and Ag/AgCl (insulin). Carrier solution, 0.1 M NaCl/0.05 M phosphate buffer (pH 7.4); flow rate, 0.5 mL/min.
Figure 1. Schematic drawing of the integrated needle-type glucose/ insulin microsensor. (A) 1, stainless steel needle body (reference electrode for glucose); 2, glucose sensor; 3, insulin sensor; 4, Ag/ AgCl (reference electrode for insulin). (B) Cross-sectional view of the sensor tip: 1, Ag/AgCl; 2, insulating layer; 3, stainless steel; 4, glucose sensor; 5, Teflon tubing wall; 6, insulin sensor.
recorders (Houston Instrument). The flow injection system consisted of a carrier reservoir, a Rainin model 5401 sample injection valve (200-µL loop), interconnecting PTFE tubing, and an Alitea C-4V peristaltic pump (Sweden). The outlet of the valve was connected to silicon rubber tubing into which the needle sensor was inserted. The currents were measured in two-electrode systems, with stainless steel and Ag/AgCl serving as reference electrodes for glucose and insulin measurements, respectively. All of the voltages are reported against these reference electrode. Fabrication of the integrated sensor. (a) Glucose enzyme microelectrode: The metallized carbon-paste glucose microelectrode was prepared by hand-mixing 10 mg of glucose oxidase with 100 mg rhodium-containing carbon paste (40% rhodium-on-carbon and 60% mineral oil, w/w). A portion of the resulting paste was packed into the end of a 7-cm-long Teflon tube (250-µm i.d., 600µm o.d.). The paste filled the tip to a height of 5 mm; electrical contact to its inner end was made with a 0.2-mm-diameter copper wire. (b) Insulin-sensitive RuOx-modified electrode: The preparation of the carbon-paste microelectrode was similar to that of the glucose biosensor. The carbon paste was obtained by hand-mixing the 70% graphite powder (nonmetalied) with 30% mineral oil (w/w) for 20 min. Prior to its modification, a carbon-paste microelectrode was pretreated electrochemically in 0.1M KNO3 medium by polarizing it at 1.8 V vs Ag/AgCl (3M KCl) for 30 s. This was followed by scanning the potential between -1.0 and 1.6 V at 100 mV/s for 10 cycles in the same solution. Subsequently, the electrode was thoroughly rinsed with water. The RuOx film was deposited electrochemically on the microelectrode surface from a 0.30 mM RuCl3/10 mM HClO4 solution by cycling its potential between -0.8 and 0.65 V at 10 V/s for 25 min. The electrode modification was initiated immediately after dissolving the RuCl3 in the HClO4 solution. The deposition was
carried out in a three-electrode system in connection to a Ag/AgCl reference electrode and a platinum wire counter electrode. The electrode was then used for insulin measurements or stored in distilled water. (c) Assembly of the Integrated Dual Sensor: The surface of a 14-gauge needle (Hamilton, Fisher) was coated with an insulation ink (Ercon) to form a thin film (∼50 µm) that was cured at 110 °C for 30 min. A Ag/AgCl layer (∼50 µm) was coated on the insulation film and was cured for 30 min in the oven. The glucose and insulin microelectrodes were inserted into the needle body. The design of the resulting dual-electrode combination sensor is shown schematically in Figure 1. Reagents. Glucose oxidase (GOD), (EC 1.1.3.4, 210 units/ mg, from Aspergillus niger), β-D(+)glucose, Bovine insulin [28 USP units, anhydrous]/mg, and ruthenium(III) chloride were purchased from Sigma Chemical. Graphite powder (Acheson grade 38) was obtained from Fisher. Rhodium-on-carbon (5% Rh. Cat. no. 20,616-4) and mineral oil were obtained from Aldrich. Other chemicals were at least ACS reagent grade and were used without further purification. RESULTS AND DISCUSSION Flow injection analysis of glucose, insulin, and their mixtures was used for assessing the temporal response and overall analytical performance of the dual microsensor. Figure 2 displays typical dual-channel flow injection current-time profiles at the needle-type glucose/insulin detector for alternate injections of 5 mM glucose (A) and 400 nM insulin (B) solutions. Both microsensors responded rapidly to injections of the corresponding target analytes, with a nearly instantaneous rise in the current, a slower decay, and peak widths of ∼90s. Such a fast response is needed for obtaining high temporal resolution. Despite the substantial difference in concentration (5 orders of magnitude) and the use of the same operating potential, there is no apparent cross reactivity. The glucose sensor (A) is, thus, not affected by injections of insulin samples while the insulin one (B) is not responding to the large excess of glucose. Notice also the favorable signal-to-noise characteristics for the submicromolar level of insulin. The optimization of the metal-dispersed carbon-paste glucose biosensor was described earlier.12 The RuOx modification was Analytical Chemistry, Vol. 73, No. 4, February 15, 2001
845
Figure 3. Simultaneous flow injection analysis of glucose/insulin mixture solutions. (A) Glucose response: concentration, 3 mM (a), 5 mM (b), and 7 mM (c). (B) Insulin response: concentration, 200 nM (a-c). Other conditions as in Figure 2.
optimized previously in connection to glassy-carbon9 or carbonfiber10 electrodes, but not in connection to carbon-paste ones. Factors influencing the performance of the insulin microsensor were, thus, examined and optimized. For example, the insulin response increased rapidly with the RuOx plating time up to 10 min, then more slowly, and leveled off above 20 min. The effect of the detection potential was assessed from the flow injection hydrodynamic voltammogram. The insulin anodic response started at +0.30 V, rose sharply to +0.55 V, and leveled off at higher values (not shown). All subsequent work, thus, employed a 25min plating time and a detection potential of +0.60 V. A similar potential (on the current plateau) was used for the biosensing of glucose. It should be pointed out that in the absence of external coatings, such potential would result in interferences from coexisting electroactive compounds. Figure 3 shows the simultaneous response of the glucose (A) and insulin (B) microsensors to mixtures containing varying levels of glucose and a fixed concentration of insulin (a-c). As desired, the glucose signal increases linearly with the concentration (over this 3-7 mM range) without influencing the response for 200 nM insulin. The latter remains stable throughout this 55-min period of continuous operation. Notice also the high reproducibility of both glucose and insulin signals for the series of 12 repetitive injections of the 7 mM glucose/200 nM insulin sample mixture (c). The mean peak currents for this series are 860 (A) and 243 (B) pA, with relative standard deviations of 2.6 and 4.1%, respectively. Similarly, the dual-channel data of Figure 4 indicate that the integrated needle sensor responds favorably, linearly, and reproducibly to increasing levels of insulin [300-1200 nM, a-c(A)] without affecting the response to 2 mM glucose (B). Overall, Figures 3 and 4 demonstrate that despite the substantial concentration difference [mM(glucose) and nM(insulin)], varying the concentration of one analyte has no affect upon the output of the second sensor. The dual insulin/glucose microsensor displays a well-defined concentration dependence. Figure 5 depicts flow-injection calibration data for insulin (A) and glucose (B) over the concentration ranges 100-1000 nM and 2-14 mM, respectively. The response peak for both of the analytes increases proportionally to the concentration to yield highly linear calibration plots [also shown as insets; sensitivity, 875 pA/µM (A); 245 pA/mM (B)]. The 846 Analytical Chemistry, Vol. 73, No. 4, February 15, 2001
Figure 4. Simultaneous flow injection analysis of glucose/insulin mixture solutions. (A) Insulin response: concentration, 300 nM (a), 600 nM (b), and1200 nM (c). (B) Glucose response: concentration, 2 mM (a-c). Other conditions as in Figure 2.
Figure 5. (A) Response to increasing levels of insulin: 100 nM(a), 200 nM(b), 300 nM (c), 500 nM (d), 700 nM (e), and 1000 nM (f). (B) Response to increasing levels of glucose: 2 mM (a), 4 mM (b), 6 mM (c), 10 mM (d), 12 mM (e), and 14 mM (f). Other conditions as in Figure 2. Insets display the resulting calibration plots.
favorable linearity of the enzyme microelectrode, obtained without an external membrane barrier, reflects the intrinsic substratediffusion barrier properties of the carbon-paste matrix. Detection limits of 50 nM insulin and 0.4 mM glucose can be estimated on the basis of the signal-to-noise characteristics of these data (S/N ) 3). Detection limits of 23 nM and 500 nM insulin were reported at the RuOx10 and RuO/CN-Ru13 coated electrodes, respectively. The stability of the dual probe was examined over a 10-day period (using the same surfaces, with intermittent storage at 4 (13) Kennedy, R. T.; Huang, L.; Atkinson, M. A.; Dush, P. Anal. Chem. 1993, 65, 1882.
°C). As expected,12 the glucose microsensor displayed a negligible decrease (of about 8%) of its sensitivity to 6 mM glucose during that period. The insulin sensor retained a highly stable signal to 400 nM insulin over the first 3 days, with a gradual daily decay (of about 4%) thereafter. Further assessment of the operational stability (of the equilibrium response) would be required in connection with in-vivo studies. In conclusion, we have demonstrated that largely differing levels of insulin and glucose can be monitored simultaneously using a needle-type combination microsensor. The results illustrate the independence of the two analytical signals that are obtained with the dual sensor. The availability of an in-vivo sensor that is capable of simultaneous monitoring of both glucose and insulin holds great promise for improved management of diabetes. Work is, thus, in progress in our laboratory for addressing the issues relevant to transforming the new dual microsensor to a truly implantable operational device. Particular attention is given to substantial miniaturization and to the shape of the device (as desired for convenient implantation and minimal disturbance of the in-vivo environment), to the use of external coatings (for
minimizing surface biofouling and redox interferences), and to the long-term stability of the equilibrium response. In view of the extensive data on implantable glucose devices,1,2 we are focusing on enhancement of the sensitivity and selectivity of the insulin probe and assessment of its reversibility and stability. Faster response time would be required for observing the temporal patterns of insulin secretion. Attention is also given to the oxygen demand of the glucose biosensor (in connection to an oxygenrich binder) and to the design of a dual-potentiostatic electronic circuitry. The multiple-analyte sensing approach can be extended to the integration of additional chemical sensors and biosensors. ACKNOWLEDGMENT This project was supported by the National Institutes of Health (NIH Grant RO1 RR14173-02).
Received for review August 9, 2000. Accepted November 29, 2000. AC0009393
Analytical Chemistry, Vol. 73, No. 4, February 15, 2001
847
