Micro-nano Matrix-based Mass Spectrometry Metabolic Fingerprint For Diagnosis And Prognosiss

In-vitro diagnostics（IVD）is an integral part of the clinics for their rapid diagnosis and prognostic assessment.Traditional IVD relies on single indicators,which are not accurate enough to meet clinical demands and is difficult to reflect the diseases process comprehensively.Metabolic fingerprint,as a branch of metabolomics,aims to analyze the variation of metabolites in organisms through qualitative and quantitative methods to elucidate the relationship between metabolic differences and physiological processes and has the potential to be applied to accurate diagnosis and comprehensive evaluation of diseases.Mass spectrometry（MS）,with its advantages of high sensitivity,high resolution,and wide detection range,is one of the main methods for metabolic fingerprint.Specially,laser desorption ionization mass spectrometry（LDI-MS）technology has great advantages in terms of ease of operation,high analytical speed,and high detection throughput,and owns the broad application prospect.However,the application of LDI-MS technology in metabolic fingerprint is greatly limited by the construction of the detection platform,including:1)the traditional organic matrix has insufficient auxiliary ionization ability in the low molecular weight range and strong background interference,which makes it difficult to accurately detect the metabolic information of clinical samples;2)For diseases involving multiple regulatory mechanisms,the accuracy of single-modal metabolic diagnosis still remains to be improved;3)large clinical cohort analysis has not achieved automated detection,and the detection efficiency of manual operation is low.Currently,the development of inorganic micro-nano matrices for LDI-MS is the focus of research,which is expected to solve the mentioned problems.Specifically,inorganic micro-nano materials with controllable components and structure can enhance the electron transition,photo-thermal conversion and energy transfer in LDI process and improve the sensitivity of LDI-MS by reasonably regulating the components and structure.Also,the rationally designed micro-nano materials can assist various detecting platforms to collect multi-dimensional data for multi-modal detection.In addition,mass spectrometry chips based on inorganic micro-nano materials can couple microarrays to automate the analytical process and significantly improve the detection efficiency.Hence,this thesis is carried out in the following three aspects:1)a metabolic fingerprint collecting and analysis process based on a novel micro-nano matrix was established to achieve highly sensitive and accurate measurement of metabolites in biological samples and successfully applied to rapid diagnosis of systemic lupus erythematosus during pregnancy;2)based on the previous work,serum metabolic fingerprints and vibrational fingerprints were collected using a dual-functional nano matrix sequentially and the bimodal data were fused to achieve highly accurate diagnosis of stroke;3)a microarray mass spectrometry chip was developed for automated detection of clinical samples to facilitate rapid diagnosis and prognosis assessments of cervical cancer.In the first part of the work,we designed hollow carbon/cobalt oxide nanocomposites（Co₃O₄/C）for aiding in the diagnosis and activity assessment of systemic lupus erythematosus（SLE）during pregnancy.Owing to the metal oxide/carbon heterogeneous structure,homogeneous surface morphology,and rough surface with nanoscale voids,Co₃O₄/C has excellent photoconversion capability,detection signal reproducibility,and small molecule selectivity.Therefore,the Co3O4/C-based LDI-MS platform can directly extract metabolic fingerprints of biosamples within seconds without pre-enrichment and purification.Combined with the optimized machine learning algorithm,we identified a unique metabolic profile consisting of four metabolic markers and constructed a simplified diagnostic panel with an AUC value of 0.9 for SLE diagnosis and activity assessment in pregnancy.This study offers the possibility of minimally invasive and rapid diagnosis and activity assessment of SLE during pregnancy in the clinicalIn the second part of the work,we constructed a rapid diagnostic platform for stroke using a bifunctional gold-palladium alloy material（Pd Au@Au CNCs）.The material features a wide localized surface plasmon resonance（LSPR）range,enhanced analyte adsorption,high selectivity,reproducibility,and stability for simultaneous high-throughput co-acquisition of surface enhanced Raman scattering（SERS）and LDI-MS data.The results demonstrated that the dual-modality SERS/LDI-MS platform based on Pd Au@Au CNCs has the advantages of simple sample pre-processing,fast data acquisition（～2 min/sample）,and low sample consumption（0.1μL serum）,which is suitable for the current stage of clinical testing.By fusing data from two different dimensions,the AUC value for stroke diagnosis reached 0.911,which is higher than the diagnostic efficiency of the two types of data alone（0.738 and 0.798,respectively）.Finally,we identified five metabolic markers and matched them with molecular vibrational features to construct a simplified stroke screening panel.This study advances the fusion of mass spectrometry and spectroscopy data for clinical diagnosis.In the third part of the work,we constructed a plasmonic LDI-MS chip for high-throughput automated microarray metabolic analysis of micro samples.Gold nanoparticles with plasmonic exciton effect are uniformly grown on the surface of regularly arranged monolayer polydopamine vesicles,forming a size-selective gap structure that enables highly selective and sensitive analysis of metabolic small molecules.Also,the analytical performance of the composite mass spectrometry chip is superior to that of the pure gold nanoparticles,and the detection sensitivity can be improved by 2 orders of magnitude.By coupling the plasmonic chip with the microarray device,we enabled the automated extraction of metabolic fingerprints of serum from cervical cancer patients.Then we successfully established a metabolic diagnostic model for the surgical prognosis of cervical cancer patients using the automated mass spectrometry platform and screened five characteristic mass spectrometry signals that can be used for rapid assessment of patient prognosis.In summary,this thesis designed and synthesized three inorganic micro-nano matrix/chip materials to establish a highly sensitive and accurate LDI-MS metabolic analysis platform.The thesis also enabled the sequential co-extraction of mass spectrometry and spectroscopy data for bimodal fusion diagnosis and explored the high-throughput automated mass spectrometry detection technology.The above results were successfully applied to the establishment of metabolic-assisted diagnosis and prognostic assessment models for SLE,stroke,and cervical cancer during pregnancy.This thesis contributes to the improvement of sensitivity,accuracy,and detection efficiency of the LDI-MS metabolic analysis-assisted diagnostic platform.Also,the results of this thesis have the potential to be applied to the adjunctive diagnosis and prognostic assessment of SLE,stroke,and cervical cancer in pregnancy in clinical settings.