11+! Methylation modification patterns based on m6A regulatory factors in colon cancer are characterized by different tumor microenvironment immune spectra

Introduction

m6A Abnormal expression and gene changes of regulatory factors are associated with malignant tumor progression and abnormal immune regulation .

Background introduction

In recent years , There have been more and more studies from m6A Study the immune related information of cancer from the perspective of regulatory factors , Today, Xiaobian brings you this article by integrating colon cancer （CC） Sample data , Assessed m6A Embellishment pattern and TME Correlation of cell infiltration characteristics . Published in 《Theranostics》 On , The influence factor is 11.556, The title of the article is ：m6A regulator-based methylation modification patterns characterized by distinct tumor microenvironment immune profiles in colon cancer.

Data is introduced

CC The gene expression data and clinical features of the samples were retrospectively collected in NCBI GEO Database and TCGA The public data set of the database .TCGA RNA Sequencing data (FPKM Format ) from UCSC Xena download , And converted to TPM Format .TCGA-COAD Genomic mutation data ( Including somatic mutations and copy number mutations ) come from UCSC Xena Database and Davoli The study of et al .

Result analysis

01

Colon cancer m6A Overview of genetic variation of regulatory factors

In this study , The author has studied 23 individual m6A RNA Methylated regulatory genes in CC The role of , Include writers、readers、erasers. The result is shown in Fig. 1A Shown , these m6A The dynamic reversible process of regulatory factors can be identified 、 Remove and add m6A Modification site , And change a lot of biological processes , Such as RNA Cut and splice 、RNA Translation and RNA The degradation . Yes 23 individual m6A The regulatory factors were analyzed GO Enrichment and meta analysis , Significant enrichment pathways are shown in the figure 1B Shown . It has been measured CC in 23 individual m6A The incidence of somatic mutations in regulatory factors , The result is shown in Fig. 1C-D Shown ,CNV Mutations are common ,23 individual m6A Of a regulator on a chromosome CNV Change the position as shown in the figure 1E Shown .

Next, the authors performed principal component analysis based on paired normal tumor samples (PCA), Find out 23 individual m6A Regulatory factors completely distinguish CC Sample and normal sample ( chart 1F). Further analysis showed that there was significant difference between up and down regulation of related genes in tumor samples （ chart 1G）.

chart 1

02

23 It is mediated by three regulatory factors m6A Identification of methylation modification patterns

23 individual m6A Interactions between regulatory factors 、 Regulatory factor linkage and its application in CC The prognostic significance of the patients is shown in the figure 2A Shown , It turns out that ,“writers”、“readers” and “erasers” The interaction among the three regulatory factors may be m6A Different modification patterns play an important role in the formation of , And it is related to the incidence and development of tumor . Therefore, the author adopts NMF The consensus clustering analysis of the algorithm is from GEO The obtained cancer samples were stratified , Three different modified aggregation zones were identified （ chart 2B）, Named as m6A-C1（221）、m6A-C2（530） and m6A-C3（162）, among m6A-C1 Life is best ,m6A-C3 The prognosis is the worst . Next, the author makes a comment on TCGA-COAD The queue performs the same analysis , Similar results were obtained ( chart 2C).

chart 2

03

m6A The modification pattern has obvious immune landscape characteristics

To explore three different m6A Biomolecular changes in modification mode , The author of Hallmarker The gene set was GSVA Enrichment analysis ( chart 2D), Results show ,m6A-C1 In the process of immune activation ,m6A-C3 Enrichment pathways are significantly associated with carcinogenic activation and matrix pathways ,m6A-C2 It is highly enriched in both immune regulation and matrix related signaling pathways .

For visualization and comparison m6A In decoration mode 28 Relative abundance of subpopulations of immune infiltrating cells , Author use ssGSEA Build a heat map ( chart 3A), Discovery of anti-tumor lymphocyte subsets , Such as effect memory CD4+/CD8+ T cells , Mainly concentrated in m6A-C1 and m6A-C2 Subtype , And in the m6A-C3 In subtypes , Regulatory T Cells and so on increased significantly . The author also uses CIBERSORT Further characterization of the immune infiltration profile , Use support vector regression to evaluate TME Subset of immune cells in , stay m6A Consistent results were also observed in the methylation modification delamination ( chart 3B).

chart 3

Next , Author use ESTIMATE Algorithm to quantify the overall immune cell infiltration of three modification patterns (immune Score) And tumor cell purity (tumor purity), Results show ,m6A-C1 The highest immune score was obtained , The second is m6A-C2 and m6A-C3( chart 3C, Upper figure ). contrary ,m6A-C3 The purity of the tumor is higher than m6A-C2 and m6A-C1, explain m6A-C2 and m6A-C1 Subtypes of tumors are characterized by more nontumor components ( Such as immune cells and stromal cells ) Surround ( chart 3C, The figure below ).

At present, some researches will CC Patients are divided into 4 Major molecular subtypes (CIN、CSC、dMMR and KRASm), This study found that CIN Subtypes are mainly concentrated in m6A-C2 and m6A-C3, and dMMR Subtypes are mainly concentrated in m6A-C1 Within the tumor ( chart 3D). It is known that PD-L1 It is predicted that PD-1/L1 Mature biomarkers of therapeutic response , The author also compares the differences m6A Decorate the cluster PD-L1 Expression level of , Found in m6A- c1 Subtypes were significantly up-regulated ( chart 3E). The above results confirm the three m6A Modification patterns have different characteristics of immune infiltration .

04

Colon cancer m6A Phenotype related DEGs

To study 3 Kind of m6A Potential gene changes and expression perturbations in modified phenotypes , This study further studies CC Three species m6A The possible existence of in a decorating pattern m6A Related transcriptional expression changes , The empirical Bayes method is applied to determine three kinds of m6A Overlapping differentially expressed genes between modification patterns (DEGs)（ chart 4A）. Of these characteristic genes GO Enrichment analysis shows , And RNA modification 、 Biological processes related to transcription and immune regulation are significantly enriched ( chart 4B).

chart 4

be based on 524 The most representative m6A Phenotype related marker genes , An unsupervised consistency cluster analysis is carried out , To obtain the 3 Stable transcriptome phenotypes . These layers divide the patient into three different m6A Gene marker subgroup , It has different clinicopathological features , And is defined as m6A gene - s1、m6A gene - s2 and m6A gene - s3( chart 4C), The authors found that patients with advanced clinical stage were m6A gene -S3 Subgroups are represented by ,CIN Subtypes and PD-L1 Patients with down-regulated expression are mainly concentrated in m6A gene - s2 and S3 Subgroup . Then the author makes a survival analysis ,m6A gene - s1 Markers have been shown to be associated with better prognosis , and m6A gene - s3 Associated with poor survival outcomes ( chart 4D), Consider age 、 Gender 、MMR And after the installment ,m6A The correlation between genetic markers and survival rate was still statistically significant （ chart 4E）.

05

m6Sig The construction of scoring system and its clinical relevance

In order to accurately predict m6A Patterns of methylation modification , The author developed a scoring scheme , be called m6Sig score , The scoring scheme is based on the identified m6A Related marker genes , To quantify individual CC Patient's m6A Embellishment mode , Use the Sange diagram ( chart 5A) Illustrates the m6Sig Workflow of scoring construction . The author passes Spearman The known biological characteristics and m6Sig The relationship between scores . The heat map of the correlation matrix shows ,m6Sig Scores were correlated with immune activation processes and DNA Repair signatures are significantly negatively correlated , But with EMT And matrix related signatures ( chart 5B).

Besides , The author also determined that m6Sig The prognostic power of the score in predicting survival outcomes , The result shows low m6Sig Scores were significantly associated with better prognosis CIT queue ( chart 5C). And the author explores m6Sig Relationship between scores and molecular subtypes , Find out dMMR Subtypes and m6Sig The score is lower than others CC Subtype ( chart 5D).PD-L1 Expression level of , stay m6Sig The low rating group was also found to have a significant increase ( chart 5E).

chart 5

combination TCGA-COAD Clinical features and genomic information of the database , The author is interested in the m6A The scoring system is verified . The study found that ,m6Sig Score on TCGA There is potential prognostic value in the cohort ,m6Sig Patients with low scores have significant survival benefits （ chart 5F).TCGA The analysis shows that CC Comprehensive molecular characteristics of , And it is suggested that according to MSI Status subdivides the tumor into 4 A subtype .MSI-H Of subtype samples m6Sig The scores were significantly lower than those of the other three subtypes ( chart 5G).

chart 5

More and more evidence shows that , The association between somatic mutations in the tumor genome and the response to immunotherapy . Therefore, the authors studied the tumor mutation load in different m6Sig The distribution pattern in the scoring group , Find out m6Sig The group with low score had higher mutation frequency ( chart 5H). The author found out in m6Sig In the low score subgroup , Changes in the copy number of somatic cells (SCNA) More frequently , namely SCNA It is positively correlated with immune evasion and tumor cell proliferation ( chart 5I). The author further points out that CC Sample m6Sig Significant mutation genes were detected in low and high groups (SMG) analysis .SMG The mutation map shows ,PIK3CA and SMAD4 stay m6Sig There was a higher somatic mutation rate in the low evaluation group , and TP53 stay m6Sig There was a higher rate of somatic mutation in the high score subtypes ( chart 5J). To learn more about CC The mutation process at work in , The author from TCGA-COAD Three mutation markers were extracted from the mutation spectrum （1,6,10）, Results found ,m6Sig The low score subtype has a significantly higher proportion of mutation characteristics 6 ( chart 5J).

06

m6Sig The role of scoring in predicting the benefits of immunotherapy

Immunotherapy is undoubtedly a major breakthrough in anti-tumor therapy , The author found ,TIDE stay m6Sig Significantly lower in the low rating group , and IPS stay m6Sig Significantly higher in the low rating group (TIDE stay TCGA-COAD and CIT Distribution in , chart 6A-B;IPS stay TCGA-COAD and CIT Distribution in , chart 6C-D). These findings indirectly prove that tumors m6A The expression of modification patterns plays an important role in mediating immune responses .

because m6Sig Strong correlation between score and immune response , The author has studied m6A Whether modifier markers can predict the response of patients in three independent immunotherapy cohorts ICI The response to treatment . Results found , In resistance PD1 Queue and anti CTLA4 In line ,m6Sig The patients in the low evaluation group showed significant clinical advantages and significantly prolonged survival ( resist PD-1, chart 6E, resist CTLA-4, chart 6G). And m6Sig Compared with patients with high scores ,m6Sig Patients with low scores confirmed ICI Significant therapeutic benefit and immune response of the treatment ( resist PD-1, chart 6F; resist CTLA-4, chart 6H). It turns out that ,m6Sig The score is related to the response to immunotherapy , It can further predict the prognosis of patients .

chart 6

Editor's summary

In this study , The author found three different m6A Methylation modification mode , They are characterized by different immunophenotypes , It is related to different anti-cancer immunity , We have also established a system called “m6Sig score ” A quantitative system to define different m6A Embellishment mode , So as to more accurately guide the treatment strategies of individual patients .