Multivariate Analysis Of Circulating Tumor Cells And Metabonomics In The Diagnosis Of Early Non-small Cell Lung Cancer

According to statistics from the International Agency for Research on Cancer(IARC)of the World Health Organization,there were approximately2.09 million new cases of lung cancer,and approximately 1.76 million deaths worldwide in 2018,with the morbidity and mortality ranking first^[1],and the5-year survival rate is only 18%.The main reason is that early lung cancer lacks typical clinical symptoms,and it is already in the middle or late stage when it is diagnosed,or has metastasized.Therefore,the key factor in reducing deaths and improving prognosis is early diagnosis of lung cancer.At present,the common methods of lung cancer diagnosis include imaging examination,sputum exfoliation cytology and bronchoscopy.These methods are easy to cause missed diagnosis and misdiagnosis,and the equipment is expensive.Therefore,it is urgent to find new biomarkers for early diagnosis of lung cancer.Traditional tumor markers of lung cancer are antigenic substances produced by tumor cells that lack or have minimal content in normal cells,and can be used for tumor screening,prognosis monitoring,and treatment effect monitoring.The common lung cancer marker carcinoembryonic antigen(CEA),cytokeratin 19 fragment(CYFRA21-1),neuron-specific-enolase(NSE),progastrin releasing peptide(Pro-GRP),squamous cell carcinoma antigen(SCC)have been widely used in the diagnosis of lung cancer,but they cannot be used for early screening.Cytokine(CK)is an important part of the immune system and participates in the normal immune homeostasis of the body.Under abnormal conditions,it can also become an important medium for the occurrence and development of some diseases,and participate in the occurrence of inflammation and tumors.Inflammatory cytokines are the bridge between chronic inflammation and lung cancer.By detecting the expression level of cytokines in the serum can provide reference value for early diagnosis of lung cancer.Circulating tumor cells(CTCs)refer to tumor cells that originate from tumor tissues and enter the peripheral blood circulation.Even in the early stages of tumor formation,a large number of tumor cells may flow into the circulatory system,so CTCs can be used as a tool of non-invasive and real-time monitoring of tumors.It is also one of the important markers of liquid biopsy of tumor patients.A French study said that CTCs can detect early lung cancer patients 4-5 years earlier than LDCT in patients with COPD.The early detection of CTCs is mainly based on simple counting.In recent years,CTCs have been sequenced by next generation sequencing(NGS),which provides a new weapon for the research of CTCs.Our previous studies have shown that LDCT combined with CTCs detection can improve the accuracy of early lung cancer screening.NGS analysis of the captured CTCs found mutations in genes such as KIT,SMARCB1,TP53,ERBB2,PDGFRA,CFS1R and FGFR1.Metabolomics is the study of life activities that are taking place at a certain moment.As a supplement to genomics and proteomics,metabolomics can perform qualitative and quantitative analysis of all metabolites in the development of tumors.Studying the expression of metabolites and the relationship between metabolites and physiological and pathological changes is helpful to screen biomarkers for early diagnosis of tumor patients.Studies have shown that the LC-MS-based metabolomics method has found 56different metabolites in the serum metabolomics of non-smoking female patients with non-small cell lung cancer.The metabolic profile is characterized by energy metabolism disorder,amino acid metabolism disorder and lipid metabolism disorder Etc.Musharraf et al analyzed the plasma of lung cancer patients,patients with chronic obstructive pulmonary disease,healthy smokers and healthy non-smokers by GC-MS,and found that plasma fatty acid and glucose levels in lung cancer patients were higher than those of other groups.The research of Ding et al.suggests that the disorder of glucose metabolism may be closely related to lung cancer.The above-mentioned studies have discovered the enhanced glucose metabolism and the weakened fat metabolism of lung cancer through metabolomics methods,and discovered some potential tumor markers of lung cancer.Therefore,it is of great significance to find biomarkers for early diagnosis of lung cancer through metabolomics.In this study,we used the peripheral blood circulating tumor cells,traditional tumor markers,and combined detection of cytokines in patients with early NSCLC to improve the diagnosis rate of early NSCLC.We used single-cell sequencing to detect the mutant genes of CTCs in early NSCLC patients,and used LC-MS the non-targeted metabonomics method qualitatively analyzes the metabolites in the serum of patients with early NSCLC and healthy controls,and analyzed and explored the gene mutation and metabolic disorder status of patients with early NSCLC,so as to distinguish healthy people from patients with early NSCLC and to explore the biological markers for early NSCLC diagnosis.Part One Significance of circulating tumor cells,traditional tumor mar-kers and combined detection of cytokines in the diagnosis of early NSCLCObjective:To explore the value of circulating tumor cells(CTCs),traditional tumor markers and combined detection of cytokines in the early diagnosis of non-small cell lung cancer.Methods:For the first time in 2018-2019,48 patients with early stage non-small cell lung cancer who were diagnosed after inpatient operation in the Thoracic Surgery Department of the Fourth Hospital of Hebei Medical University were used as the case group,and 50 healthy physical examination subjects were used as the control group.Detect the circulating tumor cell count,tumor markers and cytokines in the peripheral blood of the two groups.1.In this experiment,the new CTCs capture device Cell Collector was used to capture and enrich CTCs in the peripheral blood of early lung cancer patients and healthy people,and immunofluorescence technology was used for detection and identification.2.Using Roche cobas e 801 automatic chemiluminescence immunoassay analyzer to detect the level of the peripheral serum tumor markers carcin-oembryonic antigen(CEA),cytokeratin 19 fragment(CYFRA21-1),neuron-specific-enolase(NSE),progastrin releasing peptide(Pro-GRP),and squamous cell carcinoma antigen(SCC)of the early lung cancer group and the healthy group.3.Using Beckman NAVIOS flow cytometer microsphere array to detect the level of the cytokines IL-2,IL-4,IL-6,IL-10,TNF-a,INF-g and IL-17A in the peripheral serum of the early lung cancer group and the healthy group.4.Using receiver operating curve(ROC)to analyze the diagnostic value of individual detection and combined detection for early lung cancer,and using logistic regression analysis to compare the sensitivity and specificity of combined detection and single detection of tumor markers.All data are statistically analyzed by SPSS 22.0 software,P<0.05 indicates that the difference is statistically significant.Results:1.The level of CTCs count in peripheral blood of patients with early NSCLC was significantly higher than that of healthy people,and there was no statistical difference in CTCs and/or PD-L1 expression between the level of CTCs count and PD-L1 expression and age,sex,smoking history,clinicopathological stage and tumor size(P>0.05).2.The ROC curve was used to analyze the effect of CTCs expression level on the diagnosis of early NSCLC.The area under the ROC curve(AUC)of CTCs for the diagnosis of early lung cancer was 0.813,the diagnostic sensitivity was 62.5%,and the expression rate of CTCs was statistically significant for the diagnosis of early NSCLC.(P<0.01)3.Separating detection of peripheral blood tumor markers in early NSCLC group,and the levels of markers CEA(p<0.001),CYFAR21-1(P<0.1),and NSE(P<0.001)were higher than those in the healthy group,and the differences were statistically significant.Among them,NSE has the highest sensitivity,and CEA and CYFAR21-1 have high specificity.4.Using the Logistic regression model to fit the combined detection results of CEA,CYFAR21-1,and NSE to establish a regression diagnosis model Logit(P).The sensitivity of the combined detection of three CEA,CYFAR21-1,and NSE tumor markers in the diagnosis of early NSCLC can reach 77.08%,the specificity is 74.00%.5.Comparing the diagnostic effect of CTCs count with that of tumor markers alone or in combination,the results showed that(CTCs+CEA)had the best diagnostic effect on early lung cancer,with sensitivity of 87.5%and specificity of 82.0%.Secondly,the combined detection of several tumor markers(CEA+CYFAR21-1+NSE)showed a sensitivity of 89.58%and a specificity of 74.00%.6.The comparison of cytokine levels in peripheral blood between early NSCLC group and healthy group showed that there were significant differences in IL-2,IL-4,IL-6,IL-10,TNF-a and IL-17A levels between early lung cancer group and healthy group.(P<0.05)Further analysis showed that IL-17A had the best diagnostic effect on early NSCLC.7.The detection of CTCs combined with IL-17A has a sensitivity of91.67%and a specificity of 73.33%in the diagnosis of early NSCLC.8.The sensitivity and specificity of CTCs combined with CEA and IL-17A in the diagnosis of early NSCLC were 95.83%and 58.00%.Conclusions:1.In this study,the new CTCs capture device Cell Collector was used to enrich circulating tumor cells in vivo,and the immunofluorescence method was used to detect circulating tumor cells,confirming that this method has high sensitivity and specificity and can be applied to the detection of early NSCLC.2.Detection of tumor markers has certain value in the diagnosis of early NSCLC.CTCs combined with traditional tumor markers can improve the detection rate of early lung cancer,and CTCs combined with CEA2 has the best diagnostic effect.3.The sensitivity and specificity of CTCs combined with cytokine IL-7Awere higher than those of single detection.4.CTCs combined detection of serum tumor markers and/or cytokines can obviously improve the diagnostic sensitivity of patients with early lung cancer,which has potential clinical value for the diagnosis of early NSCLC.Part Two Second generation sequencing of gene mutations in peripheral blood circulating tumor cells of patients with early non-small cell lung cancerObjective:Using NGS to detect CTCs in the peripheral blood of 19patients with early NSCLC qualified for quality inspection,analyze the gene mutations of CTCs cells,and exploring the relationship between CTCs gene mutations and the occurrence and development of early lung cancer,so as to provide corresponding molecular markers for screening and diagnosis of early NSCLC.Methods:1.After the collection of CTCs in peripheral blood of early lung cancer is completed,stain and identify the cell collector of the captured CTCs according to the instructions of the staining kit,and provide a negative control(NK92cells)and a positive control(SK-BR-3 cells).And carry out CD45(Exbio,clone Mem-28-Alexa647)antibody,cytokeratin 7/19/pan-CK antibody(Exbio Praha,clone A53-B/A2-Alexa488)and PD-L1(clone PD-L1,Abcam)antibody Staining analysis,followed by nuclear staining with Hoechst33342(Sigma)to identify tumor cells and analyze the expression of PD-L1 in CTCs in patients with early NSCLC.2.After identifying CTCs by immunofluorescence staining,locate and cut CTCs area under microscope.A small number of CTCs from sampling needles were collected into PCR tubes,and use MALBAC method to perform whole-genome amplification.The quantitative analysis was carried out by Qubit3.0 and Nanodrop2000(Thermo Fisher).Results:1.The second-generation sequencing of CTCs in peripheral blood of 19patients with early NSCLC detected 127 gene mutations,and 94.7%of patients with early NSCLC had more than 50%of CTCs gene mutations.Among them,4 patients had peripheral blood CTCs gene mutation ratio>70%,12 patients had peripheral blood CTCs gene mutation ratio between 60%and 70%,and 2patients had peripheral blood CTCs gene mutation ratio between 50%and 60%.And in 1 case,the mutation of CTCs gene in peripheral blood was less than50%.2.Analysis of CTCs gene mutation frequency in peripheral blood of 19patients with early NSNLC showed that the high frequency mutation genes of CTCs in peripheral blood of 19 patients with early NSCLC were NOTCH1,IGF2,EGFR,PTCH1 and ARID1A.3.By NGS of peripheral blood CTCs in 19 patients with early NSNLC showed that a total of 62 mutations of lung cancer-related genes and 229mutation sites were detected.Through statistical analysis of high-frequency mutation genes,it is found that TP53 and ARID1A are the most frequent multi-site mutations.4.The results of NGS showed that TP53,NOTCH1,ARID1A,SMARCA4gene mutation frequency was higher in peripheral blood CTCs gene mutation of patients with early NSCLC,accounting for 37.99%of all mutation sites.5.The second generation sequencing results showed that in the mutation of CTCs gene in peripheral blood of patients with early NSCLC,TP53 gene mutations were co-mutated with ARID1A in 11 cases,with NOTCH1 in 11cases,with SMARCA4 gene in 9 cases,and with ALK in 8 cases.TP53 gene mutation and EGFR,ALK,CDKN2A and other 60 genes have multiple gene mutations.Conclusion:the second-generation sequencing results showed that there were 62 lung cancer-related mutations in peripheral blood CTCs gene mutations of 19 patients with early NSCLC,among which TP53,NOTCH1,ARID1A and SMARCA4 were high-frequency mutations associated with peripheral blood CTCs gene of early lung cancer,and whether these gene mutations can be used as molecular markers for early NSCLC screening needs further confirmation.Part Three The application of metabonomics multivariate analysis in early diagnosis of NSCLCObjective:To study the difference of metabolites between early lung cancer group and healthy group,and to explore the value of LC-MS non-targeted metabolomics in early diagnosis of lung cancer.Methods:1.Clinical sample collection:sera from early NSCLC patients and healthy controls were collected,including 30 early lung cancer CTCs-positive patients(all stage I and II adenocarcinoma)and 30 healthy controls,and fasting blood samples were taken.Collect 5 ml of whole blood into a sterile coagulated BD vacuum blood collection tube,immediately mix it upside down5-8 times,at 30-120 minutes,4 C and centrifuge at 4 degrees Celsius for 10minutes,transfer 1300 g and 0.2-1 ml of serum To a 1.5 ml EP tube,store at-80°C.At the same time,30 healthy volunteers took fasting blood samples,and the preparation of serum was the same as that of lung cancer group.2.Used LC-MS non-targeted metabonomics method to detect early NSCLC and control serum for analysis.Collect 100?L of serum sample,and use 400?L methanol:acetonitrile(1:1,v/v)solution to extract metabolites.The mixture was vortexed for 30 seconds and ultrasonicated on ice for 10 minutes,and this step was repeated 3 times.The samples were left at-20? for 30minutes.After centrifugation at 13000 g for 15 minutes at 4°C,the supernatant was carefully transferred to a sample bottle for LC-MS/MS analysis.3.Using Uplc-Triple-Tof-Ms platform(AB SCIEX,USA)to analyze metabolites.Exion LCTMAD system(AB Sciex,USA),ACQUITY UPLC BEH C18 column(100 mm 2.1 mm inner diameter,1.7?m;water area,Milford,USA)was used.The mobile phase A is water(containing 0.1%formic acid),the mobile phase B is acetonitrile/isopropanol(1/1)(containing 0.1%formic acid);the flow rate is 0.40 m L/min,the injection volume is 20?L,and the column temperature is 40.As part of the system adjustment and quality control process,a composite quality control sample(QC)is prepared by mixing all samples of equal volume.QC samples are injected regularly(every9 samples)to monitor the stability of analysis.The QC samples are treated and tested in the same way as the analytical samples.It is best to represent the whole sample set and monitor the stability of the analysis.4.After UPLC-TOF/MS analysis,the raw data was imported into Proge-nesis QI 2.3(Waters Corporation,Milford,USA)for peak detection and comparison.The result of preprocessing generates a data matrix composed of retention time(RT),mass-to-charge ratio(m/z)value and peak intensity.At least 50%of the metabolic characteristics of all samples are retained.After filtering,estimate half of the lowest metabolite value for a particular metabolite.In these specific samples,the metabolite level is lower than the lower limit of quantification,and each metabolite feature is normalized by sum^[19].Use QC samples for data quality control(repeatability),delete the relative standard deviation(RSD)variable>30%of quality control samples to obtain the final data matrix for subsequent analysis.5.Non-target metabolomics statistical analysis,through matching with the database(http://www.hmdb.ca/,https://metlin.scripps.edu/),finally get the metabolite list and data matrix.Principal component analysis(PCA)and partial least squares discriminant analysis(PLS-DA),orthogonal partial least squares discriminant analysis(OPLS-DA)were used to analyze differential metabolites.The effectiveness of the model is evaluated by the model parameters R2 and Q2 which provide information for the interpretability and predictability of the model,and avoid the risk of overfitting.The OPLS-DA model was used to calculate the importance of predictor variables(VIP),and the analysis of differential metabolites was performed.The screening criteria were that the importance of variable projection(VIP)was greater than 1.0 and the P value of the t test result was<0.05.The database used for metabolite-related pathway analysis is the Kyoto Encyclopedia of Genes and Genomes(KEGG;http://www.genome.jp/kegg/).6.Using SPSS 22.0 software to analyze the data.The measurement data used the mean and standard deviation to obey the normal distribution,the median(quartile)was used for the non-normal distribution,and the count data used the frequency or rate.Using the t test to compare the measurement data of normal distribution,and using the rank sum test to compare the measurement data of non-normal distribution.Using the chi-square test to compare count data.Using the ROC curve to analyze the diagnostic effect of different indicators on lung cancer.It is assumed that P-values<0.05 is significant.Results:1.The serum samples of early NSCLC group(LC)and healthy group(NC)were scanned by non-target LC-MS detection,and the original data and maps were obtained.The 1,212 metabolites of all main metabolic pathways covered were detected and analyzed,covering 31 main metabolic pathways.2.Performed PCA analysis on the data,and observed the overall metabolic difference between early NSCLC group and healthy group and the difference degree between samples in the group.The PCA scores showed that there were obvious differences in metabolites between early lung cancer group and healthy group.Used Ropls Version1.6.2 software PCA to analyze the trend of differential metabolite data between 30 early NSCLC samples and healthy group samples,and the variation between early NSCLC patients and healthy controls was observed.PCA,PLS-DA,OPLS-DA scores show that early NSCLC patients can be clearly separated from healthy group samples.Therefore,it is considered that there are obvious differences in individual metabolites between early NSCLC group and healthy group.3.Statistical analysis found that there were 100 metabolites with different content between the groups.Application of PLS-DA,Student's t-test and Fold Analysis(FC)in ropls Version1.6.2 software can distinguish the differential metabolic characteristics of the early NSCLC group and the healthy group.Combining VIP>1,P<0.05 and FC>1 or FC<1,it was found that there were100 metabolic characteristics with statistically significant differences between the two groups of samples.The most significant up-regulation of differential metabolites in the early stage of NSCLC involves lipid substances,especially sphingolipids(such as trihexylceramide)and glycerophospholipids(such as cardiolipin),which are components of cell membranes,while differe-ntial metabolites are most significantly down-regulated.Cyclic guanosine phos-phate and guanosine 1-phosphate play an important role in purine nucleo-tide metabolism.In short,we believe that these 100 different metabolites are closely related to early NSCLC.4.The 100 differential metabolites selected were classified as compounds by the Human Metabolome Database(HMDB),and pathway annotation and enrichment analysis were performed using the KEGG(Kyoto Encyclopedia of Genes and Genomes)database.We screened out 6 different types of metabolites,and KEGG pathway analysis is divided into six categories.The three most important pathways enriched by KEGG are choline metabolism,glycerophospholipid metabolism and retrograde endocannabinoid signaling pathways in cancer.5.The occurrence and development of tumors are closely related to glucose and lipid metabolism.In this study,through the analysis of the KEGG pathway,there were many different metabolites enriched in the glycolipid metabolism pathway and the tumor metabolism pathway.Among them,21differential metabolites involved lipid metabolism pathways,5 differential metabolites involved polysaccharide synthesis and metabolic pathways,and15 differential metabolites involved tumor metabolism pathways.Those were the pathway we focus on.6.Screening the potential biomarkers of early NSCLC group,we further compared the differential metabolites in lipid metabolism pathway,polysaccharide synthesis and metabolism pathway and tumor-related pathway between early NSCLC group and healthy group,and obtained 10 differential metabolites by using workers'test curve(ROC)and area under curve(AUC),of which 3 were up-regulated and 7 were down-regulated.Conclusions:1.This study explored the method of non-target metabolomics to study markers for early diagnosis of NSCLC.This method carries out semi-quantitative analysis on the markers,so that the screened markers have repeatability and effectively improve their clinical application value.2.Early-stage NSCLC patients showed different metabolic profile characteristics from healthy people.The content of multiple metabolites in early-stage NSCLC patients was significantly different from that of healthy people.This provides a certain reference basis for further experiments and clinical research,and for the realization of early NSCLC diagnosis provides certain guidance.3.Through our research,10 different metabolites as biomarkers of potential early non-small cell lung cancer can be used for auxiliary diagnosis,and the results need to be further verified by targeted metabolomics technology.4.Multivariate analysis of metabonomics showed that differential metabolites were mainly concentrated in glucose and lipid metabolism pathway,especially phospholipid metabolism pathway,and lipid metabolism played an important role in the occurrence and development of early non-small cell lung cancer.We suggest that the key enzymes in this pathway may be used as new therapeutic targets for future research.