Microfluidic Approach For Simultaneous Investigation Of The Osmotic Responses Of Multiple Oocytes

The cryopreservation of oocytes plays a significant role in the fertility preservation and the construction of oocyte bank.However,the success rate of oocytes is still very low,due to the suboptimal preservation processes such as the addition/removal of the cryoprotective agents and the speed of cooling/rewarming.The temperature-dependent oocyte membrane permeability(Lp,Ps)can provide theoretical guidance for the optimization of cryopreservation.However,those systems for studying the temperature dependence of oocyte membrane permeability are either too complicated in the microfabrication or unable to achieve a wide-range precise temperature control.Moreover,those systems cannot achieve the simultaneous observation of multiple oocytes,so that the experiment is inefficient.In addition,the data processing to acquire the cell volume uses either the time-consuming manual processing method,or some traditional image processing methods with complex steps,bad robustness and low accuracy.In order to solve these problems,this thesis dosed the following research:Firstly,a microfluidic device was designed and fabricated for the permeability study of oocytes.The device used the indium tin oxide(ITO)conductive film to form a microheater and temperature microsensor,to achieve a more accurate local temperature control compared with the global temperature control.The additional independent water bath module realized the temperature environment below room temperature,without increasing the fabrication difficulty.The ingenious combination of a local heater/sensor and a global water bath achieved relatively simple microfabrication,high accuracy and wide-range temperature control at the same time.The device could capture four oocytes with non-interfering.The captured position is controllable and the success rate of capture is high.The permeability response of four oocytes was simultaneously observed and recorded under the microscope at the same time,which improved the efficiency of the experiments.The reasonable hardware circuits,software system and anti-integral saturation PID algorithm were designed to realize an accurate temperature control of the device.In addition,the rationality and reliability of the microfluidic platform were comprehensively verified by multiphysics simulation,temperature control performance test and determination of the solution concentration change in the microchannel.Furthermore,we determined the volume excursion of the mouse oocytes exposed in cryoprotective agents(1.5 M EG and 1.5 M PG)at four temperatures(4,15,25 and 37 ℃)using the proposed microfluidic platform in this thesis.During the data processing,the deep learning of convolutional neural network was introduced into the cell permeability study for image segmentation.4000,500,and 500 pairs of training set images,validation set images,and test set images,respectively,were used for the neural network training.The accuracy of the validation set and the test set achieved 0.9860 and 0.9846,respectively.The training curve also showed that there is no overfitting for the neural network.The network completed the data processing of a large number of images with extremely high speed while a high accuracy is ensured,which greatly saved the data processing time.Finally,the temperature-dependent permeability was obtained by fitting the oocyte volume response.The results were consistent with the mouse oocyte permeability parameters determined by previous studies,indicating that the microfluidic platform with temperature control,multi-oocytes capture and the deep learning image processing method was practical,accurate and reliable.In conclusion,this study is conducive to improving the determination of oocyte membrane permeability,and further favorable for optimizing the oocyte cryopreservation methods to achieve a higher success rate of cryopreservation.