Development of an Insulin Sensor

Abstract

Every year a significant proportion of the population of New Zealand, and other countries world-wide, is diagnosed with diabetes. People suffering from diabetes have elevated blood glucose levels, which may lead to complications such as kidney failure or blindness. Diabetes is managed via the indirect measurement of the blood glucose concentration and subsequent calculations of the required insulin dose. These calculations are based on models and experience. As a result, chances are that an incorrect amount of insulin may be injected. Measuring the insulin and blood glucose concentration simultaneously would minimise the risk of injecting the wrong amount by enabling a better estimate of the required insulin dose. To improve therapy, this thesis reports the development of a sensor to detect insulin, which is to be ultimately used in a needle-free point-of-care (PoC) device.

Initially, a gold surface was modified with insulin specific aptamers. The aptamers’ change in conformation in the presence of insulin can be utilized to measure insulin concentration. Aptamers were labelled with methylene blue (MB) as a redox indicator with the goal of measuring the change in the current due to the presence of insulin. Initially, the effectiveness of different cleaning methods of gold surfaces was tested via water contact angle (WCA) in order to prepare a reproducible surface to modify. Surfaces with the highest reproducible measured WCA were trialled for modification with aptamers and 6-mercapto-1-hexanol (MCH). Aptamer- modification led to a decrease in WCAs for three sample sets, while the same could not be concluded for the modification with MCH as no significant change in WCAs was observed. Alternating current voltammetry (ACV) measurements, using the surfaces modified with aptamers and MCH as working electrodes, were performed and the peak currents in the absence and presence of insulin were compared. Peak currents of measurements without insulin were taken as jP,0, and ∆jP/jP,0 values were calculated for each measurement. Ten sensors showed the expected decrease in ∆jP/jP,0 with insulin concentration. However, no sensor set showed reproducible results, making this system unlikely to be successful as a PoC device. Based on these results, a different sensor platform, a carbon nanotube (CNT) field-effect transistor (FET) as transducer was chosen.

CNT FETs were fabricated using silicon wafers, with a SiO2 layer grown on the surface. After dicing, CNTs were deposited onto the oxide layer. To define the channel dimensions, the CNTs were etched away, and gold contacts were deposited. Finally, a passivation layer was applied. The fabrication process was optimised, including deposition of CNTs, influence of mercaptopyridine linker, and etching of CNTs. CNT density, the amount of CNTs on the SiO2 surface, as a function of deposition time, was analysed via atomic force microscopy (AFM) and the density calculated using the Mipar software. With increasing deposition time the CNT density increased, however, agglomeration also increased, leading to a point at which the forces between the CNTs are stronger than the adhesion to the SiO2 surface and a maximum CNT density. For channel definition, etching of CNTs using plasma ashing with different etching times was analysed using AFM. All CNTs were removed after 40 s. Following this, the interaction between CNTs, 1-pyrenebutanoic acid succinimidyl ester (PBASE) and aptamers was analysed via transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS). TEM images showed CNT features, a rigid and well-ordered structure. When modified with PBASE a film formed over the CNTs. After modifying with aptamers, coiled, long, thin structures could be observed. While it was concluded that aptamers were present on the surface,it could not be determined whether aptamers were adsorbed onto CNTs or bound via PBASE. XPS analysis showed an on-average higher nitrogen concentration for samples modified with CNTs, PBASE and aptamers, which indicated the presence of nucleic bases. Peaks of XPS measurements could be assigned to PBASE and to aptamers. A sample holder was built in order to perform electrical measurements, however, problems with accidental penetration of the SiO2 layer of the CNT FETs meant alternative interfacing options had to be investigated. Probe stations were used to measure the back gating of the transistors. Measurements performed with samples fabricated and measured in Wellington showed different behaviour when measured in Christchurch. Reasons for this could not be fully established despite significant efforts and in the interest of time, it was decided to switch to a more reliable system.

To this end, highly oriented pyrolytic graphite (HOPG) electrodes were modified with aptamers using PBASE as linker, and non-specific binding of insulin and other blood proteins was eliminated using either bovine serum albumin (BSA) or Tween 20. The different modification steps were analysed using cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS). Modification steps showed an increase in RCT, with the exception of some electrodes after aptamer-modification. Measurements in the absence and presence of insulin were performed and analysed using EIS. Calculated charge transfer resistance of solutions not containing insulin was taken as RCT,0, and for each insulin solution ∆RCT/RCT,0 was calculated. Ions and contaminates bound to the HOPG surface over time and non-specific interactions of insulin occurred with the electrodes. For electrodes modified with blocking agent alone, BSA showed the smallest change in ∆RCT/RCT,0 over time, whereas electrodes modified with Tween 20 showed slight improvement. An electrode modified with PBASE, aptamers and BSA, showed a steep increase in RCT at 10 fM, which could be attributed to the interaction of insulin with aptamers. In contrast, no evidence could be found that electrodes modified with PBASE, aptamers and Tween 20 showed response of aptamers wrapping around insulin.