Study of Robot-aided Microscope System with 3D Image Reconstruction Information for Precise Intracellular Manipulation

Abstract

Intracellular manipulation assisted by a robotic microscope system has valuable applications in genetic diagnosis and genome-editing tasks. However, current intracellular manipulation suffers from a low success rate and large operation damage owing to insufficient information on the 3D positions of selected organelles. In this thesis, an image reconstruction method using a robot-aided microscope system and its applications in intracellular surgery is studied. 3D position information can be obtained through the 3D reconstruction of organelles and can be fed to an automatic manipulator, thereby improving success rate while reducing operation damage. This research is conducted based on the following three aspects:

First, an image reconstruction method with 0.1 µm resolution for carrying intracellular manipulation based on optical tomography of wide-field fluorescence microscope (WFFM) is proposed. The objective lens moves along the z axis with step size for the scanning of specimens. Sampled sections are deconvoluted by a noise regulated maximum likelihood estimation (NRMLE) deconvolution algorithm using multiple point spread functions (PSFs) at each sampled plane. Then, a filtered region- based segmentation method is utilized to obtain the clear boundary of each deconvoluted section for 3D model reconstruction. Simulation tests are performed on a microsphere in 3D space at different noise levels to demonstrate the accuracy of the proposed 3D reconstruction method as well as the advantages this method over other deconvolution methods using fixed and approximated PSFs. The results of reconstruction tests on a 4 µm fluorescent microsphere and intracellular nuclei of human dermal fibroblasts, neonatal (HDFn) cells extract authentic PSFs of the developed microscope system and prove that the proposed method is accurate enough to be applied for 3D information feedback of intracellular operations.

Second, an intracellular cell delivery technology with 3D reconstruction information is developed on HDFn cells. A 3D model of the targeted HDFn nucleus is reconstructed, and an articulated 3D geometry analysis is developed for the selection of an optimal delivery position. A robust controller that can compensate for external disturbances is utilized to manipulate micropipettes to deliver exogenous materials. Nucleus delivery experiments without the use of 3D information are conducted for comparison. The delivery success rate of the proposed method is 71% while that of 2D method is 48%. The nucleus delivery results with 3D reconstruction information exhibit great advantages in delivery precision and efficiency.

Third, the developed robot-aided cell manipulation system is employed for mitochondria extraction experiments on human leukemia monocytic (THP-1) cells. The 3D reconstruction model of the intracellular mitochondrion is generated first. Given the low thickness of a mitochondrion (<2 µm ), the optimal extraction position is selected in a plane containing the largest section area of the reconstructed model. Mitochondria extraction experiments are conducted and compared with the approach with 2D information. The extraction successful rate of the proposed method is 60% while that of 2D method is 40%. These results show that the proposed method greatly improves the extraction success rate and reduces operation damage on cells.

In summary, this study presents a 3D image reconstruction method based on a robot-aided WFFM system and intracellular manipulation applications using 3D reconstruction information. This thesis represents a significant progress in the development of precisely targeted intracellular manipulation. The success of this study will constitute a technological platform for cell-based therapy and genetic diagnosis for precise medicine in the future.

Date of Award	6 Feb 2020
Original language	English
Awarding Institution	City University of Hong Kong
Supervisor	Dong SUN (Supervisor) & Zhuangde Jiang (External Supervisor)