| Currently,cancer is the first killer of human beings,and lung cancer has the highest morbidity and mortality.So the research on lung cancer has great significance.VOCs in human exhaled breathe is an ideal non-invasive markers for lung cancer screening.However,due to the lack of research between biological machenism and lung cancer,it is difficult to determine the con:fident specific VOCs.In addition,the government spends a lot of money to develope new drugs for lung cancer every year,and the development of each drug needs a long process.There are still many shortcomings in current drug screening,so it is necessary to develop microscopic and high-throughput drug screening technologies.Overall,lung cancer diagnosis model was optimized and the relationship between VOC and lung cancer was verified.Various drug screening models of lung cancer were also cultured and drug tests were conducted on impedance sensors for 3D spherical cells.Finally,preliminary experiment of lung-on-a-chip was carried out.The main research of this thesis is as follows:1.Lung cancer diagnosis model based on VOCs was optimized by adding four variables.The input variables of the model are optimized by random forest,Thirty-one VOCs were screened from human exhaled breathe by Receiver Operator Characteristic Curve.Lung cancer diagnosis model has good discriminatory ability between LC group and HC group,but not well between LC group and PNMD group.Therefore,four variables of gender,age,smoking status and CT were added into the previous model,and the sensitivity,specificity and overall accuracy were significantly improved.In order to reduce the input variables of the model,we used the random forest algorithm to establish a lung cancer diagnosis model with five variables,and the accuracy was same as previous model.It contributes to the application of lung cancer diagnosis in clinical practice in the future.2.The relationship between VOC in exhaled breath and metabolic pathways in lung cancer patients was explained through bioinformatics analysis.Transcriptome data of lung cancer patients and healthy people in the TCGA database was analyzed by R language algorithm to screen differentially expressed genes.Then,KOBAS 2.0 enrichment analysis was used to annotate these genes and locate to the relevant metabolic pathways in vivo.The results showed that there were significant differences in gene expression between lung cancer patients and healthy people.Through the comparison with the metabolic pathway database of VOC production,it was found that the DE genes were consistent with human VOC metabolic pathways.It means that these VOC markers were correlated with the metabolism of lung cancer patients.In addition,genetic mutation analysis and protein structure analysis were used to explain the possible mechanism of VOC production.3.Three drug screening models for lung cancer including patient-derived tissue,3D cells and organoid were developed.Cell from lung cancer patients can survive for a week after seeding on a petri dish.Nude mice experiment was used to proliferation of PDTC for high-throughput drug screening.In vitro 3D cell can mimic the physiological state and the surrounding growth environment in human body,and good results have been obtained in 3D culture with or without scaffolds.At present,organoid is one of the most promising drug screening models.It is similar to a certain organ or tissue in structure and function level.We have cultured various organoids like intestine,olfaction and lungs.By combining microfluidic and micromanufacture technologies,organoid-on-a-chip will have great contribute to drug screening in the future.4.Bionic 3D spheroids biosensor chips based on impedance sensors for high-throughput and dynamic drug screeningLung cancer 3D spheroids models cultured by A549 and H1299 cell lines were constructed by suspension drop method.The cell state of 3D spheroids was detected by electrical impedance sensor in real-time way.Cisplatin was used to detect the drug effect between 2D cells and 3D spheroids.The results showed that 2D cells had a stronger efficacy to cisplatin,and that cisplatin had a stronger effect to H1299 cell line.In addition,two combination drugs were performed and cisplatin plus etoposide have stronger drug efficacy.Multi-organs-on-a-chip platform was developed to simulate the interaction between different organs in human body for drug screening.5.The progress in the fabrication and application of organ-on-a-chip was reviewed and the preliminary research of bionic lung-on-a-chip was introduced.Humam-on-a-chip plays an important role in disease research,drug development,toxicity prediction and personalized medicine.Combineing sensor for real-time is the future development detection.Currently,electrical impedance sensors,ME A and LAPS sensors have been combined with organ-on-a-chip to detect various physiological states of cell culture.We conducted real-time detection of hydrogen ion metabolism in lung-on-a-chip and obtained corresponding drug reactions.However,only detection of hydrogen ion cannot explain the complex cell state,so nore parameters should be added to achieve accurate monitoring of cell state. |
PDF Full Text Request