Lab-on-a-Chip Integration of Size-based Separation Techniques for Isolation of Bacteria from Blood

Abstract

Clinical sample preparation is an essential process in modern diagnostics for maximizing sensitivity and specificity of detection and for ensuring reliability of assay readout. In general, sample preparation typically involves isolating and concentrating a population of target molecules, cells, or particles together with the removal of undesired components from specimen that could otherwise interfere with target detection. The identification of bacteria from complex clinical matrices such as blood presents a particular sample preparation challenge. Conventional culture-based methods typically require at least 24 h of incubation time, making this approach unsuitable for use in rapid diagnostics. Therefore, the development of sample preparation methods for bacteria with rapid processing time, high purification efficiency, and large volumetric throughput to enable analysis of low bacteria concentrations in blood remains a key challenge. This dissertation is focused on realizing a universal platform for preparing microbial sample from blood that is free lysis buffer, electric field, or affinity-based capture methods. First, we developed the porous silica monolith elements integrated into thermoplastic devices for isolation of intact bacteria from blood, enabling the application of emerging detection methods that supports bacterial identification from purified cell populations. Second, to support high throughput analysis of blood samples procured in resource-limited environments, microfluidics elements integrated directly into a syringe are demonstrated by utilizing the deterministic lateral displacement technique and the Dean flow focusing methods. Through these approaches blood cell reduction prior to bacteria isolation can be achieved, thereby increasing the overall sample volume that may be processed by the system. Additionally, a miniaturized hydrocylone capable of operating at tens of milliliters per minute feed rate is presented. Complex microstructures successfully realized at a hundred-micron scale by 3D printing technique presented a promising route to the unconventional microfluidic systems. Lastly, we demonstrated ancillary microfluidic components required to enable full operation of the system in a low-cost lab-on-a-chip format suitable for implementation in resource-limited environments and optimize overall operation of the platform to achieve throughput, sensitivity, and selectivity suitable for clinical application when coupling the platform with downstream detection methods designed for assay readout from intact bacteria.