Biocompatibility of subcutaneous implanted glucose sensors: Working toward modeling the sensor/tissue interface

Abstract

The worldwide incidence of diabetes is growing rapidly. Therefore, the impetus to provide tight insulin therapy to reduce complications aided by continuous glucose monitoring is great. After nearly three decades of research, only three continuous glucose monitoring systems are near commercial availability. The development of a reliable glucose sensor for use in these systems has been hampered by anomalies in in vivo sensor performance due to the inflammatory response. Investigation of the biocompatibility of implanted glucose sensors with respect to the inflammatory processes was undertaken. The glucose sensor function and the processes of the inflammatory response are not independent (consumption of glucose and oxygen, and production of hydrogen peroxide by both cells and sensors). However, this investigation has attempted to separate these processes which occur in the complex in vivo environment. The initial events upon implantation, which is infiltration of biomolecules or ‘biofouling’, cause an initial loss of sensitivity. In addition, the host response to the glucose sensor may lead to sensor response anomalies due to perturbation of analyte availability. Mediation of the inflammatory response was explored in order to provide a glucose sensor with improved biocompatibility. The identification of the biomolecules using proteomic techniques, with MALDI-TOF MS, revealed that fragments of whole proteins were the primary source of biofouling. Measurement of the flux components (oxygen, glucose, and hydrogen peroxide) released by inflammatory cells indicated that cellular oxygen consumption may cause significant oxygen depletion near an implanted sensor. The incorporation of nitric oxide (NO) release from a glucose sensor effectively mediated the inflammatory response while accurately monitoring in vivo glycemia. As a result of the flux and NO release sensor studies it was apparent that NO released from inflammatory cells may also contribute to sensor response fluctuations. The development of a mathematical model to study the temporal stages of the inflammatory response between the inflamed tissue and an active glucose sensor is proposed. These studies, while raising additional investigative paths, contributed key insights to the biocompatibility issues encountered for successful development of implantable glucose sensors.