Mechanical characterization and in vivo operation of an implantable drug delivery MEMS device

Abstract

The goal of this thesis was to advance an implantable drug delivery MEMS (MicroElectroMechanical Systems) device developed in our laboratory. This device was designed to locally deliver multiple substances in complex release profiles in order to maximize the effectiveness of drug therapies. It consists of an array of microreservoirs etched into a silicon substrate. Different types and dosages of drugs can be contained in these reservoirs capped by thin gold membranes. The drug release is achieved by the application of a small anodic potential on the gold membrane in a chloride containing medium (such as the body fluid). The gold membrane will corrode and disintegrate so that the drug contained within the reservoir is free to diffuse into the surrounding medium. Previous researchers have demonstrated in vitro and in vivo release of tracer molecules as well as a radiolabled chemotherapeutic agent (carmustine, or BCNU) from the device. However, systematic characterization of the mechanical and electrochemical behavior of gold membranes on the drug delivery device was necessary in order to achieve more reliable device performance and to demonstrate efficacy of BCNU delivered from the MEMS device against an experimental tumor model. A bulge test apparatus was constructed to characterize the mechanical properties of gold membranes. Uniform pressure was applied from underneath the gold membrane and the membrane deflection was measured using optical interferometry. Analyzing the deflection and pressure data allowed extraction of the elastic modulus and residual stress of the gold membrane.

(cont.) Gold membranes with in-plane sizes ranging from 20 to 200pim showed lower modulus (126-168 GPa) than bulk (111) single crystal gold (189 GPa). But their yield strength (317-351 MPa) was higher than the bulk value. An in situ experimental setup was constructed to observe the electrochemical disintegration process of the gold membranes. Real time images recorded from a CCD camera showed non-uniform corrosion occurring first around the membrane edges. Bulge tests on the corroded membranes indicated a gradual loss of mechanical integrity of the gold membranes due to corrosion. The gold membrane disintegration probably occurred by a combination of membrane thinning through active dissolution and accumulation of plastic deformation due to the transient formation of a passive film on top of the gold membrane in each voltammetry cycle. Dense gold membranes with reproducible opening behavior are critical to the success of large scale in vivo studies and future commercial applications. Defects in the gold membranes led to premature leakage of BCNU, a small molecule drug. Wafers with sputtered gold membranes patterned by wet etching had a higher device yield and membrane quality than wafers with evaporated gold membranes patterned by lift off. The mechanical and electrochemical studies provided guidance to improve the operation reliability and reproducibility of the drug delivery device. In vivo release of BCNU from the drug delivery device was demonstrated in a rat flank model. Acute temporal release kinetics of 14C labeled BCNU in vivo was evaluated by analysis of the plasma 14C concentration using the accelerator mass spectrometry (AMS) technique.

(cont.)The in vivo 14C labeled BCNU release profile from the activated devices was similar to that of the in vitro and subcutaneously injected controls. The time to reach a steady-state plasma 14C concentration was on the order of one hour. Efficacy of BCNU delivered from the drug delivery device was demonstrated in a 9L rat flank tumor model. Co-formulation of PEG with BCNU led to complete and rapid release of payload in vivo. The retarding effect of BCNU on the tumor growth was dose dependent in the range of 0.67 - 2 mg. BCNU delivered from activated devices seemed to be as effective as equipotent injections of BCNU against the tumor growth. This tumor effect study provided preliminary efficacy validation of the drug delivery device as well as important dosage information for further efficacy evaluation of the BCNU/IL-2 combination therapy.