Electrospun membranes for implantable glucose biosensors

Abstract

The goal for this thesis was to apply electrospun biomimetic coatings on implantable glucose biosensors and test their efficacy as mass-transport limiting and tissue engineering membranes, with special focus on achieving reliable and long sensing life-time for biosensors when implanted in the body. The 3D structure of electrospun membranes provides the unique combination of extensively interconnected pores, large pore volumes and mechanical strength, which are anticipated to improving sensor sensitivity. Their structure also mimics the 3D architecture of natural extracellular matrix (ECM), which is exploited to engineer tissue responses to implants. A versatile vertical electrospinning setup was built in our workshop and used to electrospin single polymer - Selectophore™ polyurethane (PU) and two polymer (coaxial) – PU and gelatin (Ge) fibre membranes. Extensive studies involving optimization of electrospinning parameters (namely solvents, polymer solution concentration, applied electric potential, polymer solution feed flow rate, distance between spinneret and collector) were carried out to obtain electrospun membranes having tailorable fibre diameters, pore sizes and thickness. The morphology (scanning electron microscopy (SEM) and optical microscopy), fibre diameter (SEM), porosity (bubble point and gravimetry methods), hydrophilicity (contact angle), solute diffusion (biodialyzer) and uniaxial mechanical properties (tensile tester) were used to characterize certain shortlisted electrospun membranes. Static and dynamic collector configurations for electrospinning fibres directly on sensor surface were optimized of which the dynamic collections system helped achieve snugly fit membranes of uniform thickness on the entire surface of the sensor. The biocompatibility and the in vivo functional efficacy of electrospun membranes off and on glucose biosensors were evaluated in rat subcutaneous implantation model. Linear increase in thickness of electrospun membranes with increasing electrospinning time was observed. Further, the smaller the fibre diameter, smaller was the pore size and higher was the fibre density (predicted), the hydrophilicity and the mechanical strength. Very thin membranes showed zero-order (Fickian diffusion exponent ‘n’ ~ 1) permeability for glucose transport. Increasing membrane thickness lowered ‘n’ value through non-Fickian towards Fickian (‘n’ = 0.5) diffusion. Thin electrospun PU membranes (~10 μm thick) did not affect, while thicknesses between 20 and 140 μm all decreased sensitivity of glucose biosensor by about 20%. PU core - Ge shell coaxial fibre membranes caused decrease in ex vivo sensitivity by up to 40%. The membranes with sub-micron to micron sized pore sizes functioned as mass-transport limiting membranes; but were not permeable to host cells when implanted in the body. However, PU-Ge coaxial fibre membranes, having <2 μm pore sizes, were infiltrated with fibroblasts and deposition of collagen in their pores. Such tissue response prevented the formation of dense fibrous capsule around the implants, which helped improve the in vivo sensor sensitivity. To conclude, this study demonstrated that electrospun membrane having tailorable fibre diameters, porosity and thickness, while having mechanical strength similar to the natural soft tissues can be spun directly on sensor surfaces. The membranes can function as mass-transport limiting membranes, while causing minimal or no effect on sensor sensitivity. With the added bioactive Ge surfaces, evidence from this study indicates that reliable long-term in vivo sensor function can be achieved.