Edmond MAGNER ? Materials and Surface Science Institute and Department of Chemical and Environmental Sciences, University of Limerick, Limerick, Ireland

Abstract:
The development and characterization of electrochemical biosensors is described from the perspective of the development of biosensors for the determination of glucose. The factors that have influenced the large scale commercial development of electrochemical glucose biosensors are discussed.

Please cite as: CHEMIK 2013, 67, 2, 105-110

Introduction

A biosensor is a device that combines a biological recognition element together with a transduction system for the detection of a specific analyte [1]. The biological component can comprise species such as a redox protein or enzyme, an antibody, a whole cell or a DNA strand and serves the purpose of imparting specificity to the sensor. The transduction system transforms the recognition event of the biological component into a measurable signal utilising optical, piezoelectric, conductimetric, thermometric, amperometric or potentiometric methods of detection. The first biosensor, for the measurement of glucose was developed by Clark in 1962 [2] and subsequently developed commercially by Yellow Springs International [3], with the first product released in 1973. Since then there has been an enormous number of reports on biosensors. A web of science search (accessed on June 19, 2012) on the topic of biosensors listed 17,663 publications. Despite this level of activity and the substantial amounts of funding and research effort, only one widely used biosensor, for the detection of glucose, has been successively commercialized on a large scale. The commercial development of biosensors, or more accurately the lack of such systems, given the amount of effort that has been expended worldwide , is a good illustration of the dogma of the development of any new form of technology which can be described in four phases; discovery, hype, cynicism and reality. Significant research funding was devoted to a broad range of areas in biosensors (discovery) with the expectation that an array of biosensors would become available on a commercial basis (hype). This level of expectation has been met for glucose testing (reality), but not for any other analyte (cynicism). This article discusses the background to the development of glucose sensors and describes the factors that need to be taken into account in the development of commercially viable sensors for other analytes.

Diabetes mellitus

The successful development of glucose biosensors can be simply ascribed to the condition of diabetes. The condition is classified into two types, type 1 is insulin dependent and usually occurs in young adults and children while type 2 is non-insulin dependent and usually occurs in obese adults over the age of 40. In both types, the body is no longer able to regulate the amount of glucose itself and external intervention in the form of quantifying the concentration of glucose together with injection of insulin is necessary to regulate blood glucose concentrations. If uncontrolled, diabetes can be a life threatening condition and is the sixth leading cause of death in the US [4].

The complications that may arise with diabetes include problems in the eyes, kidneys and nervous system together with circulatory problems. In 1993, the landmark Diabetes Control and Complications Trial (DCCT) demonstrated that ?intensive therapy effectively delays the onset and slows the progression of diabetic retinopathy, nephropathy, and neuropathy in patients with insulin dependent diabetes mellitus? [5]. An intrinsic component of such therapy is the measurement of the concentration of glucose a number of times per day. As part of the outcomes of the study, the DCCT recommended that diabetics should monitor their blood glucose levels at least four times per day. The requirement for such a measurement regime is the basis for the successful development of glucose biosensors. It has been estimated that health conditions related to diabetes account for 14% of the all of the health care costs in the United States. Given that the number of people diagnosed with diabetes is rising rapidly and is expected to become a much more significant problem in countries such as India and China, there is a clear and demonstrable need for these devices. In addition to an adequate market need and size, testing for glucose also meets the other criteria essential to the successful development of a sensor; a cheap and stable biorecognition element, a reasonably high analyte concentration and the ability to mass produce sensors on a scale of millions per day.

Detection of glucose

The concentration of glucose can be determined easily using HPLC and is usually performed in this manner (as with other sugars) in fermentation systems etc. Such methods clearly do not lend themselves to the type of testing required for a diabetic patient. The only alternative method relies on enzymatic analysis (equations (1) and (2)) by the enzyme glucose oxidase (GOx) which contains the co-factor, flavin adenine dinucleotide (FAD). The oxidized form of the enzyme (GOx(FAD)) catalyses the oxidation of glucose and is itself reduced in the process. Subsequently, the reduced form of the enzyme (GOx(FADH2) is oxidized by O2 to regenerate the oxidized form of GOx.

In addition to the requirements described above, the successful development of a system for the detection of glucose for home use has the additional requirements of:
? using an undiluted whole blood sample of volume less than 1 ml
? a disposable, single use system with a stable shelf life of greater than 18 months at room temperature
? a sensor with a stable calibration
? no requirement for additional reagents
? low cost and
? ease of use.

From equations (1) and (2), the concentration of glucose can in principle be determined by measuring the amounts of either GOx(FADH2), gluconic acid, oxygen or hydrogen peroxide. In practise the measurement of changes in pH which arise from the production of gluconic acid is not feasible due to the larger buffering capacity of blood, while changes in the concentration of O2 are not discernible due to the large amounts of O2 already present in blood. The concentration of glucose can then only be monitored by determining the concentration of GOx(FADH2) or of H2O2. Measurement of the reduced form of glucose oxidase is complicated by the fact that the flavin group is buried deeply within the amino acid matrix, rendering electron transfer to and from the redox active site and the electrode difficult. In contrast the concentration of H2O2 can be performed in a relatively straight forward manner at a platinum electrode, albeit at a relatively high potential of 0.6 V.

Detection of H2O2

The instrument developed by Yellow Springs International (YSI) determines the concentration of glucose by detecting the concentration of hydrogen peroxide (equation (2)). The instrument released in 1973 has been successively refined and is still manufactured and sold using the same measurement format (albeit with significant levels of refinement and improvement). The instrument utilises a 25 ml sample of blood which is added to 0.5 ml of buffer. The glucose concentration is then measured using a membrane covered platinum electrode. The membrane is a composite of three materials is used for the immobilisation of the enzyme and to protect the electrode from fouling by cells and proteins. Diffusion of glucose occurs through the initial, polycarbonate component of the membrane which excludes proteins and cells. The next part of the membrane contains immobilised glucose oxidase which oxidises glucose to gluconic acid, producing the reduced form of the enzyme. The enzyme is regenerated by reaction with oxygen to produce hydrogen peroxide, which then diffuses though a third, cellulose acetate layer, to the platinum electrode where it is oxidised to oxygen on application of a potential of 0.6 V. The cellulose acetate layer excludes interfering species such as ascorbate and uric acid. The sensor is robust, accurate and precise and is used in clinical laboratories as a standard method for the determination of glucose in blood samples. A feature of the sensor is that its response is not significantly affected by the haematocrit content (red cell count) of the sample. However the system is expensive and not easily miniaturised. Due to both cost and manufacturing concerns, it is not feasible to utilise a membrane such as that of the YSI instrument on a disposable electrode. In addition the instrument is not suitable for home use and in particular, is too large for ambulatory use (dimensions of ca. 30 x 20 x 15 cm).

Glucose biosensors for home use

Glucose test systems have been available for home use since the 1950?s when Ames, Inc. released urine test strips. However these systemswere not successful as the test itself was relatively complex, utilised urine samples and required accurate timing by the patient. With these systems, the timing of the test had to be controlled by the patient, leading to significant errors in the results. The concentration of glucose was determined by comparing the colour change on the strip with a standard series of colour changes. The method was prone to error, in particular given that diabetics are more prone to eye conditions and do not necessarily possess the motor coordination skills of healthy patients of the same age. Spectrophotometric based sensors were subsequently developed and released in the late 1980?s. Their principle of operation relies on using an enzyme such as glucose oxidase to oxidise glucose. The hydrogen peroxide produced is subsequently reduced by the enzyme horseradish peroxidase which in turn oxidises a dye. The colour change produced is detected on a hand held spectrophotometric system. For some time, these systems dominated the home glucose testing market.. However, spectrophotometric detectors require insertion of the test strip into the detector and require the patient to clean the detector on a regular basis (approximately once a week). As described above, such a requirement runs opposite to one of the key requirements of a successful biosensor, namely ease of use. Electrochemical based biosensors now dominate the market as the sensors for such devices are cheaper to manufacture. In 1987, MediSense, Inc. released the first electrochemically based glucose biosensor for home use [6]. The sensor was based on research [7] performed at the University of Oxford by Hill and co-workers and utilised ferrocene as a mediator to shuttle electrons from the buried flavin group of GOx to the surface of a carbon electrode. A key feature of this device was that it relied on the process of screen printing to produce the electrodes. This process was well established, low cost and amenable to the mass production of sensors. Subsequently, other commercially available, electrochemical sensors which relied on using ferricyanide as mediator were developed by Bayer [8] and Boehringer-Mannheim [9].

The performance of these sensors was reasonable for home use but did not meet the levels of accuracy and precision of clinical instruments. The advantages that these tests offer in comparison to larger, more accurate laboratory based tests lie in their ease of use and their portability, which enables diabetics to test their blood glucose levels as part of their normal daily routine. While significant improvements and refinements of the sensors have been implemented, the accuracy and precision of these devices has yet to match that of, for example, the YSI instrument. The main difficulty with these sensors is that they display a bias in the results which is a function of the haematocrit level of the samples. In addition to fouling of the electrodes by the red cells which reduces the measured current by a combination of reduced electrode area and increased solution viscosity, the red cells themselves have an effect on the response due to the excluded volume of the cells [10]. This effect is exacerbated by the presence of glucose within this volume which is not directly accessible to the enzyme on the electrode surface. Recent improvements to this type of sensor have included the use of osmium based mediators which are more effective electron shuttles and the use of high concentrations of enzyme and mediator to oxidise essentially all of the glucose in the sample [11]. Such near complete oxidation of glucose in the serum portion of the blood sample causes the rapid facilitated diffusion of glucose from within the red cell to the serum part of the sample and helps ensure close to complete oxidation of all of the glucose in the sample. More recent advances in glucose sensors lie in the development of implantable devices capable of determining the concentration of glucose, with the results of the analysis fed to a controller which can then automatically determine the amount of insulin required at that particular time and deliver the insulin via an implanted pump. The development of such devices represents the culmination of decades of effort to develop an artificial pancreas, which would enable diabetics to counteract much of the debilitating features associated with the condition [11]. To date, such devices have been show to perform at reasonable levels of accuracy for periods of a few days but suffer from deteriorations in response over longer times due to fouling, adhesion, etc.

Summary

Glucose biosensors have been successfully developed mainly because a large section of the population suffers from diabetes. The prevalence of diabetes forms the basis for a large market with a considerable financial incentive to develop and market home use blood glucose tests. In addition to an adequate market need and size, testing for glucose also meets the other criteria essential to the successful development of a sensor which include: a cheap and stable biorecognition element; a reasonably high analyte concentration; the ability to mass produce sensors on a scale of millions per day; a disposable, single use system with a stable shelf life of greater than 18 months at room temperature; small sample volumes; no requirement for additional reagents; low cost and ease of use. It has been estimated that the successful development of a glucose biosensor requires funding of the order of $100 million. Given such costs, the development of biosensors for other analytes can only succeed if the criteria listed above can be met in order to fund such high initial costs.

Literature
1. Magner, E. Trends in electrochemical biosensors, Analyst, 1998, 123, 1967-1970.
2. Clark, L.C., Lyons, C. Electrode systems for continuous ,onitoring in cardivascular surgery, Annals New York Academy of Sciences, 1962, 102, 29-35.
3. Yellow Springs Instruments Co. Inc., OH 45387, USA.
4. Nathan D.M., Cleary, P.A., Backlund, J.Y., Genuth, C.S.M., Lachin, J.M., Orchard, T.J., Raskin, P., Zinman, B. Intensive diabetes treatment and cardiovascular disease in patients with type 1 diabetes, New England Journal of Medicine, 2005, 353, 2643-2653.
5. H. Shamoon, et al., The effect of intensive treatment of diabetes on the development and progression of long-term complications in insulin-dependent diabetes mellitus. New England Journal of Medicine, 1993, 329, 977-986.
6. Higgins, I.J., Hill, H.A.O., Plotkin, E.V. US Pat. 4545382, (1985).
7. Cass, A.E. , Davis, G., Francis, G.D., Hill, H.A.O., Aston, W.J., Higgins, I.J., Plotkin, E.V., Scott, L.D. Turner, A.P. Ferrocene-mediated enzyme electrode for amperometric determination of glucose, Analytical Chemistry, 1984, 56, 667-671.
8. Nankai, S., Kawaguri, M., Iijima, T., 1990, Patent USA 4897173.
9. Pollman, K.H., Gerber, M.T., 1994, Patent USA 5288636.
10. Magner, E. Detection of ferricyanide as a probe for the effect of hematocrit in whole blood biosensors, Analyst, 126 (2001) 861-865.
11. Heller, A., Feldman, B., Electrochemistry in diabetes management, Accounts of Chemical Research 2010, 43, 963-973.

 Edmond MAGNER ? studied at University College,Cork, Ireland (B.Sc.) and the University of Rochester, New York (Ph.D.). After postdoctoral work at Imperial College, London and the Massachusetts Institute of Technology, he worked on the development of electrochemical biosensors at MediSense and Abbott Laboratories in Massachusetts. He returned to Ireland in1997 to take up an academic position at the University of Limerick. He has published 65 journal papers and over 100 conference proceedings. His current research interests are in bioelectrochemistry and biocatalysis.
e-mail address: ; phone: 353 61 202629