Research On Precise Diagnosis And Regulatory Network Of Lymphoma Based On Non-coding RNA

Lymphoma is a systemic malignancy derived from the lymphatic system,accounting for 3% to 4% of all malignant tumors.Currently,more than 40 lymphoma subtypes are defined,which can be roughly divided into Hodgkin lymphoma(HL)and Non-Hodgkin lymphoma(NHL)subtypes according to pathological characteristics.The successful treatment of lymphoma begins with an accurate diagnosis.And the existing diagnostic techniques are mainly based on tissue biopsy of cell morphology,immunophenotype and molecular diagnosis.Benign reactive lymphoid hyperplasia(RLH)is a common biopsy finding,and it often be confused with malignant lymphoma.Thus,it is not an easy task to distinguish benign RLH from lymphoma just relying on existing diagnostic methods.Therefore,it is imperative to develop diagnostic method with novel biomarker to accurately diagnose RLH and lymphoma.At the same time,diffuse large B cell lymphoma(DLBCL),as the main type of NHL,is a clinical heterogeneous disease.Although the majority of DLBCL patients can be cured with the standard R-CHOP chemotherapy,approximately 30 to 40 % of patients will develop relapsed or refractory disease.Therefore,new biomarkers are urgently needed to stratify DLBCL patients with poor prognosis.In complex organisms,60-70% of genome is transcribed into non-coding RNA(ncRNA),which has many different classes including microRNA(miRNA),small nucleolar RNA,and long non-coding RNA(lncRNA)and so on.They can regulate gene expression at various levels in the physiological and developmental processes,and their dysregulations are closely related to the occurrence and development of diseases.It indicates that they have potential to be new biomarkers for disease diagnosis and prognosis.The competing endogenous RNA(ceRNA)theory indicates that lncRNA can indirectly regulate mRNA level through sponging miRNA.It provides a new perspective for studying the pathogenesis and targeted therapy of diseases.The purpose of this thesis is to screen miRNA biomarkers and construct a classifier which can distinguish lymphoma from RLH to provide an aid for accurate diagnosis of lymphoma.At the same time,the R-CHOP efficacy classifier for DLBCL was built to stratify patients with DLBCL into prognostic subgroup.Furthermore,DLBCL and HL-specific ceRNA networks were systematically analyzed and compared.It clarified the ceRNA regulatory relationships between lncRNAs and mRNAs in these two lymphoma subtypes,and provided new candidates for the mechanism analysis and targeted therapy of DLBCL and HL.The main research contents and results are as follows:1.A 12-miRNA panel were screened to distinguish RLH and lymphoma through the analysis of next generation sequencing data and literature retrieval.After optimization of the panel,a 6-miRNA panel were constructed,among which miR-21,miR-146 a,miR-155,miR-17,miR-150 were biomarkers,and let-7f was internal reference.2.A classifier were developed by machine learning method for distinguishing lymphoma from RLH,which is based on the 6-miRNA panel.Evaluated by independent sample set including 375 RLH and 262 lymphoma samples,the accuracy was 91.37%,the sensitivity was 93.13%,the specificity was 90.13%,and the area under the curve was 0.97.It indicated that the classifier had excellent and stable performance,and can serve as a powerful method for accurately diagnosing lymphoma.3.A classifier based on miR-150,miR-146 a,miR-155 and miR-17 was established to evaluate the therapeutic effect of R-CHOP in patients with DLBCL of activated B cell subtype.With five-fold cross validation,the accuracy,sensitivity and specificity was reached 92.5%,93.13% and 90.13%,respectively.It indicated that the four miRNAs may act as biomarkers to evaluate the efficacy of R-CHOP in patients with DLBCL.4.Differential expressed lncRNAs,mRNAs and miRNAs in DLBCL and HL were identified by the differential expression analysis of microarray datasets.Based on ceRNA network theory,the DLBCL and HL-specific ceRNA networks were constructed.Moreover,the ceRNA relationships shared by the two networks were selected to form a common ceRNA sub-network.The RNAs in the sub-network may be involved in the common ceRNA regulation pathways in both lymphoma subtypes.5.In the DLBCL-specific ceRNA network,ten mRNAs were identified which were significantly related to the overall survival of DLBCL patients,including ZNF3,PPP1R2,RABEP1,PEX11 B,SIK1,KXD1,TSG101,TLE4,CDV3,STX16.A survival risk model of DLBCL was proposed using eight of these mRNAs through Lasso regression.And it has the potential to predict the survival of DLBCL patients.To sum up,this thesis focused on the diagnosis and identification,prognostic stratification and internal regulatory relationship of lymphoma based on ncRNA.It provides a new method for the accurate diagnosis of lymphoma and lays the foundation for analyzing the ncRNA regulatory network of DLBCL and HL.It has profound influence for basic research and clinical practice for lymphoma.