The Effect of Porous Poly-L-Lactic Acid Coatings on Tissue Response and Subsequent Glucose Sensor Performance

Abstract

<p>Efforts to create a reliable, long–term implantable glucose sensor have been stymied by the effects of the foreign body response and wound healing that introduce delayed response times as well as unpredictable sensor performance. Loss of vascularization from fibrotic encapsulation around implanted sensors is purported as a key contributor to sensor failure, as glucose and oxygen transport to the sensor becomes impeded. Improving sensor performance by increasing angiogenesis and/or reducing capsule thickness using tissue-modifying textured coatings is attractive because texturing is not dependent upon a depletable drug reservoir. A significant range of materials and pore sizes are capable of promoting angiogenesis and reducing capsule thickness, provided pores have open-architecture with dimensions sufficiently large enough to allow inflammatory cell infiltration. </p><p><br></p><p>Poly–L–lactic acid was gas foamed/salt leached with ammonium bicarbonate to produce porous coatings for Medtronic MiniMed SOF–sensor glucose sensors. Coating properties included 30μm pore diameters, 90% porosity, and 50μm wall thickness. Cytotoxicity, degradation, and sensor response time studies were performed to ensure the porous coatings were non–toxic and negligibly retarded glucose diffusion prior to <italic>in vivo</italic> testing. Histology was used to evaluate angiogenesis and collagen deposition adjacent to porous coated and bare (i.e. smooth, uncoated) non–functional sensor strips after three weeks in the rat dorsal subcutis. Functional Medtronic glucose sensors, with and without porous coatings, were percutaneously implanted in the rat dorsum to assess if the angiogenic–inducing properties observed around the non–functional porous coated sensor strips translated into stable, non–attenuated sensor signals over two and three weeks. MiniLink<super>TM </super>transmitters were attached to the rats, permitting continuous glucose monitoring. Vessel counts and collagen deposition adjacent to sensors were determined from histological analysis. A one–sided dorsal window model was developed to further evaluate the interplay between vascularization and sensor performance Sensors were inserted beneath the windows, allowing visualization of microvascular changes adjacent to sensor surfaces, with simultaneous evaluation of how vascular changes impacted interstitial glucose monitoring. </p><p><br></p><p>Porous coating did have angiogenic–inducing effects on the surrounding tissue. When fully implanted in the rat dorsum, sensor strips with porous coatings induced three–fold more vessels within 100μm<super>2</super> of the sensor strip surface after three weeks and two-fold more cumulative vessel lengths within 1mm<super>2</super> after two weeks, compared to bare surfaces. In contrast, when percutaneously implanted in the rat dorsum, porous coated and bare sensors were equally highly vascularized, with two–fold more vessels than fully implanted bare sensors. </p><p><br></p><p>Despite increased angiogenesis adjacent to percutaneous sensors, sensor signal attenuation occurred over 14 days, suggesting that angiogenesis plays a secondary role in maintaining sensor function. Percutaneously implanted porous coated sensors had greater reductions in baseline current (20 to 50+%) over two weeks than bare sensors (10 to 30%). Mechanical stresses imposed by percutaneous tethering may override the beneficial effects of porous coatings. Furthermore, integration of the porous coating with the surrounding tissue may have increased tissue tearing at the porous coating–tissue, increasing inflammation and collagen deposition resulting in greater signal attenuation compared with bare sensors. Future investigations of the role mechanical irritation has on wound healing around percutaneous glucose sensors are warranted.</p>