Advancing 3D volumetric imaging with computational optics for in vitro tissue model

Abstract

In vitro 3D tissue model is superseding 2D cell culture model because it has shown to replicate in vivo tissue level functions processes down to single-cell level. Accurate quantification of the 3D spatial distribution of physical or molecular markers within in vitro 3D tissue model identifies cell and tissue characteristics that are subsequently used to screen for drug targets. Fluorescence techniques equipped to acquire 3D volumetric images such as light sheet and lightfield computational imaging have emerged as promising candidates to fulfil low phototoxic volumetric image-based screening of in vitro 3D tissue model at single-cell resolution. However, these 3D volumetric imaging techniques are laden with optical and practical challenges that lead to limited axial sectioning along with sample and mechanical interventions. These obstacles in turn fail to accelerate further innovative progress in applying 3D volumetric imaging in drug discovery. The aim of this thesis is to achieve isotropic 3D volumetric imaging with computational optics so that it is applicable for high-throughput screening of in vitro 3D tissue model. In Chapter 1, I first provide an overview of different classes of in vitro 3D tissue model and the hurdles in implementing light sheet fluorescence microscopy (LFSM) for volumetric image-based screening. Next in Chapter 2, I discuss the foundations of 3D volumetric imaging by providing the theoretical background of axial sectioning, reviewing the optical techniques for voxel generation in 3D, and outlining how to achieve isotropic spatial resolution. In Chapter 3, I provide detailed biochemical protocols for the fabrication of in vitro thrombosis model and the fabrication of in vitro 3D stromal tissue model using fluid viscosity modifiers. In Chapter 4, I attempt to address the challenge of refractive index (RI) difference faced by LFSM for 3D volumetric imaging. I formulate an RI matching protocol with sucrose solution and tunable illumination to enable LFSM to image single platelets in in vitro thrombosis model. Whilst the technique developed for LFSM in Chapter 4 is suited for one class of in vitro 3D tissue model: tissue-on-chips, it is still not broadly applicable to all in vitro 3D tissue model such as spherical tissue model in multi-well plates. Hence, in Chapter 5, I move to propose a novel technique that combines single objective light sheet (SOLS) and lightfield computational imaging to achieve multi-directional SOLS. Multi-directional SOLS achieves flexible 3D volumetric imaging because it can accommodate planar illumination of many different oblique and azimuthal angles. At the core of the multi-directional SOLS technique is the Fast Optical Ray lightfield reconstruction algorithm because the FOR algorithm facilitates 3D volumetric imaging with axial sectioning that reaches near-isotropic spatial resolution. Chapter 5 describes the numerical modelling and simulation of the FOR algorithm and multi-directional SOLS. Based on the technique proposed in Chapter 5, I then detail the optical design and construction to create the multi-directional SOLS system, termed the mcSOLS in Chapter 6. Chapter 6 ends with a preliminary demonstration of the mcSOLS for 3D volumetric imaging of in vitro 3D stromal tissue model that was fabricated in Chapter 3, where I show the single-cell resolution of 3.59x3.63x4.56um in a volume of 270x270x40um. In Chapter 7, I push the capabilities of the mcSOLS beyond imaging and into photolithography. I investigate, numerically and experimentally, the feasibility of implementing the multi-directional planar illumination of the mcSOLS for tomographic photolithography. Lastly, I conclude in Chapter 8 on how the development of 3D volumetric imaging techniques and in vitro 3D tissue model in this thesis exemplifies the synergy between optics, biology, and fabrication for advancing volumetric image-based screening. This paves the way for the next generation of drug discovery platforms. Show more