ELSEVIER
Contents lists available at ScienceDirect
Current Problems in Cancer
journal homepage: www.elsevier.com/locate/cpcancer
Current Problems in Cancer
General Oncology
Comprehensive Pan-Cancer Analysis of MTF2 Effects on Human Tumors
Check for updates
Cui Tanga,Ť, Ye Lvb,t, Kuihu Dinga,1, Yu Caoª, Zemei Maa, Lina Yanga,d, Qiqi Zhangª, Haiyang Zhouª, Yu Wanga, Zhongtao Liu“, ** , Xiangmei Caoa,*
a Department of Pathology, Basic Medical School, Ningxia Medical University, Yinchuan, China
b Department of Oncology, General Hospital of Ningxia Medical University, Yinchuan, China
” Department of Neurosurgery, General Hospital of Ningxia Medical University, Yinchuan, China
d Department of Pathology, Ningxia People’s Hospital, Yinchuan, China
ABSTRACT
Understanding oncogenic processes and underlying mechanisms to advance research into human tumors is critical for effective treatment. Studies have shown that Metal regulatory transcription factor 2(MTF2) drives malignant progression in liver cancer and glioma. However, no systematic pan-cancer analysis of MTF2 has been performed. Here, we use University of California Santa Cruz, Cancer Genome Atlas , Genotype-Tissue Expression data, Tumor Immune Estimation Resource, and Clinical Proteomic Tumor Analysis Consortium bioinformatics tools to explore differential expression of MTF2 across different tu- mor types. MTF2 was found to be highly expressed in the cancer lines that were available through the respective databases included in the study, and overexpression of MTF2 may lead to a poor prognosis in tumor patients such as glioblastoma multiforme, brain lower grade glioma, KIPAN, LIHC, adrenocorti- cal carcinoma, etc. We also validated MTF2 mutations in cancer, compared MTF2 methylation levels in normal and primary tumor tissues, analyzed the association of MTF2 with the immune microenviron- ment, and validated the functional role of MTF2 in glioma U87 and U251 and breast cancer MDA-MB-231
* Conflicts of interest: All authors declare that they have no conflicts of interest.
** Ethical statement: The study did not involve animal testing, so there are no ethical claims.
* Correspondence to: Xiangmei Cao, Department of Pathology, Ningxia Medical University, 1160 Shenli South St, Yinchuan, Ningxia Hui Autonomous Region, 750004, China.
** Correspondence to: Zhongtao Liu, Department of Neurosurgery, General Hospital of Ningxia Medical University, Yinchuan, China.
E-mail addresses: L.6288@163.com (Z. Liu), caoxm.nxmu@163.com (X. Cao).
* There authors have contributed equally to this work and share first authorship.
cell lines by cytometry. This also indicates that MTF2 has a promising application prospect in cancer treatment.
@ 2023 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
ARTICLE INFO
Keywords: MTF2; Cancer; Prognosis; Methylation; Immune infiltration
Introduction
At present, the incidence of cancer is gradually increasing as the burden on society in- creases.1 Although many treatments have been clinically successful, such as surgery, im- munotherapy, radiotherapy, chemotherapy, and targeted therapies, there is still a high mortal- ity rate and many questions remain about the pathogenesis of cancer and treatment options.1,2 Therefore, it is important to explore new tumor markers for clinical cancer treatment. Currently, we can better understand malignant tumor progression by mining pan-oncogenes. There are a large number of resource databases that can be used to collect tumor-associated datasets such as The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression (GTEx) for more detailed downstream pan-cancer analysis.3-5
Polycomb group (PcG) proteins are known as epigenetic transcription inhibitors and play a key role in maintaining cells’ memory of fateful decisions made during development.6 The PcG protein is distributed in 2 major complexes, polycomb repressive complex 1/2 (ie, PRC1/2), which are responsible for H2AK119ub1 labeling maintenance, and its core components are RING1A/B.7 PRC2 is responsible for global labeling of H3K27me3, including 3 core proteins: SUZ12, EED, and 1 histone methyltransferase, EZH1 or EZH2.8 Metal regulatory transcription factor 2 (MTF2), also known as polycomb-like protein 2 (PCL2), is a catalytic inactive PCL family protein that has been identified for recruitment of PRC2 as a locus from embryonic stem cell.9,10 MTF2 has also been shown to inactivate chromosomes, which in turn regulate the activity of specific developmental genes PRC2.11 There is increasing evidence that MTF2 plays an important role in tumor progres- sion, such as strong epithelial mesenchymal transition (EMT) capabilities such as mobility and aggressiveness in most cancer cells.12 EMT is carefully choreographed by transcription factors including Twist, Snail, and the Zeb family.13 Furthermore, studies in hepatocellular carcinoma (HCC) have shown that upregulation of MTF2 expression promotes EMT.14 In addition, MTF2 is associated with multiple myeloma, acute myeloid leukemia (AML) and retinoblastoma (RB).15-17 However, previous studies on PCL2 have been limited to a few cancer types, and its impact on other tumor types is unclear.
In this study, we discussed MTF2 expression profiles in pan-cancer and analyzed MTF2 expression and clinical characteristics using a variety of bioinformatics tools, taking into ac- count genetic alterations, prognosis, protein methylation, immune infiltration, protein cross- fertilization, and cytological validation. Comprehensive analysis reveals MTF2 expression patterns in patients with metastatic cancer, screening for tumor types with poor prognosis, and provides a basis for the study of therapeutic targets.
Materials and Methods
Materials
The glioblastoma multiforme (GBM) cell lines (U87, U251) and BRCA cell lines (MDA-MB- 231) were purchased from Wuhan PromoCell Life Science Co., Ltd. (Wuhan, China). Dulbecco’s
modified eagle medium and fetal bovine serum (FBS) were purchased from Biological Indus- tries (Israel). The rLV-hMTF2-3flag-ZsGreen-Puro and rLV-shRNA2-Puro-hMTF2 were synthesized by Life Technologies (USA), and Lipofectamine 2000 was purchased from Invitrogen (Shanghai, China). Bicinchoninic acid (BCA) was purchased from Keygen (Jiangsu, China). HemAtoxylin-eosin (HE) kits was purchased from Solarbio (USA) and 24-well plate transwell chamber systems from Corning (USA).The rabbit antibody against humanEZH2 (#5246), B-actin (#5125) as well as a horseradish peroxidase-linked goat antirabbit antibody (#7074), was purchased from cell sig- naling technology; rabbit antibodies against human MTF2 (ab262915), vimentin (ab92547), and N-Cadherin (ab18203) were purchased from abcam.
Methods
Gene Expression Analysis
Open resource platform using SangerBox (http://vip.sangerbox.com/home.html), which was also used in subsequent analyses. We downloaded the unified standardized pan-cancer dataset from the University of California Santa Cruz (UCSC) (https://xenabrowser.net/) database: TCGA PanCancer (PANCAN, N = 10535, G = 60499) and TCGA TARGET GTEx (PANCAN, N = 60499, G = 60599). Further, we extracted MTF2 gene expression data from each sample, performed a log2 (x + 0.001) transformation for each expression value, and finally eliminated tumors with fewer than 3 samples in a single cancer cell line, resulting in 26 and 34 tumor lines (See Supplemental Table 1 and Table 2 for raw data). To calculate the difference in expres- sion between normal and tumor samples per tumor type, we used R software (version 3.6.4) to calculate the difference in expression between normal and tumor samples per tumor, us- ing unpaired Wilcoxon Rank Sum and Signature Rank Tests. In addition, UALCAN (http://ualcan. path.uab.edu/analysis-port.html) open resource platform was used to obtain the Clinical Pro- teomic Tumor Analysis Consortium (CPTAC) dataset for MTF2 protein expression analysis in HCC cases.
Survival Analysis
Then we extracted MTF2 gene expression data in various samples. In addition, we obtained high quality TCGA prognostic datasets from TCGA prognostic studies previously published in Cell and excluded samples with follow-up of less than 30 days,18 further transforming each expres- sion value with log2 (x + 0.001) , and finally eliminating tumors with fewer than 10 samples in a single cell line, resulting in expression data for 39 cancer lines and Overall survival (OS) data for corresponding samples (see Supplemental Table 3 for raw data). We used the coxph function of the R package survival (version 3.2-7) to establish the cox proportional hazards progression model to analyze gene expression in relation to prognosis in each tumor, and performed statisti- cal assays using Logrant test to obtain prognostic significance. MTF2 gene expression and tumor prognosis were also analyzed using the KM-plotter (https://kmplot.com/analysis/) database in the SangerBox Open Resource Platform, and optimal MTF2 cutoff values were calculated using the R package maxstat. Patients were divided into high and low groups based on a minimum sample size greater than 25% and a maximum sample size smaller than 75%, resulting in an optimal cut-off value. Prognosis differences between the 2 groups were further analyzed using the R package survivit function, and the prognostic differences between the different groups were assessed using the logrank test method. In addition, in gene mutation analysis, cBioPor- tal (https://www.cbioportal.org/) tool was used to analyze the prognostic relationship between mutational status and progression-free survival, OS, disease-free survival, and disease-specific survival in individual tumors.
Gene Clinical Analysis
Further, we extracted MTF2 gene expression data from each sample, performed a log2 (x + 0.001) transformation for each expression value, and finally eliminated tumors with fewer ☒ than 3 samples in a single cancer cell line, resulting in stage expression data for 30 cancer lines (see Supplemental Table 4 for Raw data), grade expression data for 14 cancer lines (see Supple- mentary Table 5 for raw data), and age expression data for 37 cancer lines (see Supplemental Table 6 for raw data). We used R software (version 3.6.4) to calculate differences in gene expres- sion and Pearson correlation across clinical stage samples in each tumor. Differential analysis between the 2 groups was performed using unpaired Student’s t-Test and analysis of variance (ANOVA) was performed on multiple groups of samples.
Gene Mutation Analysis
Changes in MTF2 protein structure frequency and mutation types (gene mutation, ampli- fication, and profound deletion) were analyzed in all TCGA tumors using the cBioPortal tool (https://www.cbioportal.org/). We also downloaded the Level 4 Simple Nucleotide Variation dataset from GDC (https://portal.gdc.cancer.gov/) for all TCGA samples processed by MuTect2 software.19 The domain information of the protein was obtained from the R package matools (version version 2.2.10).
RNA-Modified Gene Analysis
Further, we extracted expression data for MTF2 and 44 marker genes for 3 types of RNA mod- ifications (m1A,10 m5C,13 m6A21) from each sample (see Supplemental Table 7 for Raw data), and we also filtered all normal samples, further transforming each expression value with log2 (x + 0.001). Next, we calculated Pearson correlation between MTF2 and the marker gene. ☒
Immunocheckpoint Analysis
We further extracted the expression of the MTF2 gene and 60 labeled genes of 2 immune checkpoint pathways, Inhibitory24 and Stimulature,34 from the literature The Immune Landscape of Cancer.20 (see Supplementary Table 8 for Raw data). Further, we filtered all normal samples and performed a log2 (x + 0.001) transformation for each expression value. Next, we calculated the Pearson correlation between MTF2 and the marker genes.
Immunocyte Analysis
Further, we extracted MTF2 gene expression data from each sample and further transformed each expression value with log2 (x + 0.001) (see Supplementary Table 9 for Raw data). In ad- dition, we extracted gene expression profiles from each tumor and mapped them to Gene Sym- bol. The Timer method (TIMER: a web server for compositive analysis of tumor-infusing immune cells21) was used to reassess each patient’s B cell, T cell CD4, T cell CD8, neutral CD8, mactrophin score, and mactrophin infiltrates each tumor based on gene expression using the R package IOBR (version 0.99.922). Pearson’s correlating immune infiltration score was calculated for each tumor using the corr.test function of the R package Psych (version 2.1.6) to determine significantly cor- related immune infiltration scores.
Immune Infiltration Analysis
Finally, we extracted MTF2 gene expression data in individual samples (see Supplementary Table 10 for Raw data). Log2 (x + 0.001) transformation was further performed for each ex- pression value. In addition, we extracted gene expression profiles from each tumor and mapped them to gene symbol. Stromal, immune, and ESTIMATE scores were calculated for each patient in each tumor based on gene expression using the R package Estimate (version 1.0.1323). Fi- nally, we obtained immune infiltration scores for 9555 tumor samples from a total of 39 tumor types. Pearson’s correlating immune infiltration scores for each tumor were calculated using the corr.test function of the R package psych (version 2.1.6) to identify significantly associated im- mune infiltration scores.
Protein-Protein Interaction Networks (PPI)
PPI analysis of MTF2 using String (http://string-db.org) database. In addition, the association of MTF2 with PPI genes was analyzed using the GEPIA2 (http://gepia2.cancer-pku.cn/# index) database.
Cell Culture
Human U87, U251, and MDA-MB-231 cell lines were cultured in complete dulbecco’s modified eagle medium containing 10% FBS, 1% penicillin and streptomycin in incubators at 37℃ and 5% CO2. The cells were cultured in aseptic petri dishes with 0.25% trypsin when the cell count reached 90%.
hMTF2 / shMTF2 Lentiviral Transfection
U87, U251, and MDA-MB-231 cell lines were resuspended in a 35 mm dish with 1 x 105 cells. Incubated at 37℃ with 5% CO2 until 50%-80%, then added a medium containing rLV-shRNA2- puro-hMTF2 (shMTF2), rLV-hMTF2-3flag-ZsGreen-Puro (hMTF2) and rLV-Puro (Vector). Continue culture for 24 hours, replacing virus-containing media with fresh media. Fluorescence expression was observed and cultured to 72 hours with Puromycin screening.
Western Blotting
U87, U251, and MDA-MB-231 cells were cultured and collected from hMTF2, shMTF2 and Vector. Then add the lysate to the cell precipitation and place the lysate on ice for 30 minutes. After complete lysis, the cells were collected at a centrifuge of 15,000 x g for 10 minutes at 4℃. Protein concentration was measured using a BCA kit. Protein samples were isolated by 10% polyacrylamide gel electrophoresis and transferred to PVDF membrane. Seal with 5% skim milk at room temperature for 2 hours, cut PVDF membrane and incubate overnight in diluted anti- body (both 1: 1000). The next day was incubated with HRP-conjugate II antibody at 37℃ at a retemperature of 30 minutes, 3 times with the cleaning membrane. Protein strips were tested using an ultrasensitive chemiluminescent kit after the membrane was washed again.
Plate Colony Assay
U87, U251, and MDA-MB-231 cells from stable hMTF2, shMTF2, and vector were cultured in a 35 mm cell culture dish (1000 cells/dish). The cells were cultured in a medium containing
10% FBS for 12 days, then carefully cleaned twice with phosphate buffer saline and fixed with 4% polyformaldehyde for 15 minutes. Next, remove the fixation solution and add the crystalline purple, then carefully wash and air-dry the stained cells for 30 minutes. Count colonies directly with the naked eye or, under a microscope, the number of spheres with more than 50 cells.
Wound Healing Assay
Use the underside of a 35 mm plate to mark the crosshairs. Next, U87 and MDA-MB-231 cells from stable hMTF2, shMTF2 and Vector were cultured in these dishes (4 x 105 cells/dish). When the cell fusion degree reaches 80%-90%, horizontal lines are drawn along the single cell layer with the tip of the pipette. The cells were washed 3 times with phosphate buffer saline to remove the isolated cells and the scratch width was photographed at different times. Scratch relative width was measured using image-ProPlus 6.0 software to reflect cell migration capacity. Repeat at least 3 times per experiment.
Transwell Assay
Cell suspension of 2.5 x 105 cells / mL was prepared in FBS free medium and added to the upper chamber of the transwell chamber and immediately placed on a 24-hole plate. The lower chamber was filled with a complete medium containing 20% FBS. Cotton swabs were used to remove cells remaining in the upper chamber 24 hours after incubation. A total of 15 minutes was fixed with 4% polyformaldehyde and 0.1% crystalline purple stained 4 hours. Then it dries naturally and is photographed.
Statistical analysis
Pearson was used to analyze the correlation between variables. The significance level was set at P < 0.05. Cytology experiments were statistically analyzed using SPSS 22.0 software. Differ- ences between groups were analyzed using single factor ANOVA and Sheffe’s post-hoc test. The data were expressed as mean ± standard deviation, with each experiment repeated 3 times. P < 0.05 was considered statistically significant.
Results
Gene Expression and Protein Analysis
First, we investigated the differential expression of MTF2 between tumors and adjacent normal tissues in TCGA database. We observed significant upregulation in 15 tumors such as cervical squamous cell carcinoma (CESC), lung adenocarcinoma (LUAD), colon adenocar- cinoma (COAD), colon adenocarcinoma/rectum adenocarcinoma esophageal carcinoma (COAD- READ), esophageal carcinoma (ESCA), stomach and esophageal carcinoma (STES), pan-kidney co- hort (KICH+KIRC+KIRP) (KIPAN), stomach adenocarcinoma (STAD), uterine corpus endometrial carcinoma (UCEC), head and neck squamous cell carcinoma (HNSC), kidney renal clear cell carci- noma (KIRC), lung squamous cell carcinoma (LUSC), liver hepatocellular carcinoma (LIHC), blad- der urothelial carcinoma (BLCA), cholangiocarcinoma (CHOL), and significant downregulation in 3 tumors such as THCA, PCG, KICH (P < 0.05, Fig 1A) (See Supplementary Table 11 for statisti- cal details). Since normal tissue samples from the TCGA database are few, we further analyzed MTF2 expression levels in combination with the GTEx database. We observed significant up- regulation in 18 tumors such as GBM, glioblastoma multiforme glioma (GBMLGG), lower grade
A
TCGA
8
Expression
6
4
7
Group
1
HE
E
İ
A
1
İH
S
I
HE
14
2
H:
F
I
₹
f
H
1
I
F
LE
H
Tumor
.
0
Normal
-2
- 4
CESC(T=304,N=3)
LUAD(T=513,N=109)>
COAD(T=288,N=41)>
COADREAD(T=380,N=51)
ESCA(T=181,N=13)
STES(T=595,N=49)
KIPAN(T=884,N= 129)>
STAD(T=414,N=36)>
UCEC(T=180,N=23)
HNSC(T=518,N=44)
KIRC(T=530,N= 129)
LUSC(T=498,N=109)>
LIHC(T=369,N=50)
THCA(T=504,N=59)
PCPG(T=177,N=3)
BLCA(T=407,N=19)
KICH(T=66,N=129)
CHOL(T=36,N=9)
B
TCGA+GTEx
15
MTF2 Expression
10
5
H
IF
4
1
HI
La
II
41
I
Ba
4
4
A
5.
4
0
4
5
1
A
1
N
ER
II
A
E
n
3
”
1
0
Group
Tumor
-5
Normal
-10
-15
GBM(T=153,N=1157)
GBMLGG(T=662,N=1157)
LGG(T=509,N=1157)
UCEC(T=180,N=23)
BRCA(T=1092,N=292)
CESC(T=304,N=13)
LUAD(T=513,N=397)
ESCA(T=181,N=668)
STES(T=595,N=879)
KIRP(T=288,N=168)
KIPAN(T=884,N= 168)
COAD(T=288,N=349)
COADREAD(T=380,N=359)
PRAD(T=495,N= 152)
STAD(T=414,N=211)
HNSC(T=518,N=44)
KIRC(T=530,N= 168)
LUSC(T=498,N=397)
LIHC(T=369,N=160)
WT(T=120,N=168)
SKCM(T=102,N=558)
BLCA(T=407,N=28)
THCA(T=504,N=338)
READ(T=92,N=10)
OV(T=419,N=88)
PAAD(T=178,N= 171)
TGCT(T=148,N=165)
UCS(T=57,N=78)
ALL(T=132,N=337)
LAML(T=173,N=337)
PCPG(T=177,N=3)
ACC(T=77,N=128)
KICH(T=66,N=168)
CHOL(T=36,N=9)
Protein expression of MTF2 in Hepatocellular carcinoma
C
2
P=8.47E-10
I
1
CPTAC+dataset
Z-value
0
“7
2
3
Normal(n=165)
Primary tumor(n=165)
CPTAC samples
glioma (LGG), UCEC, ESCA, STES, KIPAN, COAD, COADREAD, STAD, HNSC, KIRC, Wilms tumor, BLCA, pancreatic adenocarcinoma (PAAD), acute lymphoblastic leukemia (ALL), LAML, CHOL, and significant downregulation in 8 tumors such as LUAD, prostate adenocarcinoma (PRAD), skin cu- taneous melanoma (SKCM), THCA, ovarian (OV), pheochromocytoma and paraganglioma (PCPG), adrenocortical carcinoma (ACC), KICH ([P < 0.05, Fig 1B] [See Supplementary Table 12 for sta- tistical details]). Expression differences between normal and tumor samples in each tumor were calculated using R software (version 3.6.4), and significant differences were analyzed using un- paired Wilcoxon Rank Sum and Signed Rank Tests. In addition, protein levels of MTF2 were eval- uated using CPTAC datasets. Protein expression in HCC was found to be significantly higher than in normal tissues (P < 0.05, Fig 1C). These results suggest that MTF2 may play an oncogenic role in various cancers.
Survival Analysis
Next, we will focus on the relationship between MTF2 expression and prognosis in each tu- mor. The results revealed that high expression levels of MTF2 were significantly associated with poorer survival in 5 patients including GBMLGG, LGG, KIPAN, LIHC, ACC, and low expression lev- els of MTF2 were significantly associated with poorer survival in 3 patients including thymoma (THYM), SKCM, and OV (P< 0.05, Fig 2A). (See Supplementary Table 13 for statistical details). We further analyzed MTF2 expression with LGG, KIPAN, LIHC, ACC, SKCM, and OV prognosis using the KM-plotter database. Finally, we observed that high expression of MTF2 in LGG, KIPAN, LIHC and ACC had a poor prognosis, and low expression of MTF2 in OV had a poor prognosis, while SKCM had no significant difference. (P< 0.05, Fig 2B-G)
Gene Expression and Clinical Stage Analysis of Cancer
Further analysis of the clinical relevance of MTF2 was performed. We used R software (ver- sion 3.6.4) to calculate differences in gene expression in samples at different clinical stages in each tumor, performed a significant difference analysis between the 2 using unpaired Student’s t-Test, and performed variance testing in multiple samples using ANOVA. Significant differences were observed in five tumor types such as breast invasive carcinoma (BRCA), KIPAN, THYM, LIHC, and ACC in clinical stage analysis (P < 0.05, Fig 3A). (See Supplementary Table 14 for statistical details). In the clinical grade analysis, we observed significant differences in seven tumors such as GBMLGG, LGG, KIPAN, HNSC, KIRC, LIHC, PAAD (P < 0.05, Fig 3B). (See Supplementary Ta- ble 15 for statistical details). To further analyze the age-related associations, we calculated their Pearson associations in each tumor using R software (version 3.6.4). We observed significant negative associations in 10 tumors, including GBMLGG, LGG, LUAD, BRCA, SARC, kidney renal papillary cell carcinoma (KIRP), KIRC, THCA, PAAD, BLCA (P < 0.05, Fig 3C-D). (See Supplemen- tary Table 16 for statistical details).
Genetic Mutation Analysis
Although mutations are not sufficient to cause cancer, the accumulation of mutations can lead to cancer. Therefore, we next focused on exploring MTF2 gene alterations in human tumor samples. According to our analysis, the highest frequency of MTF2 alterations was in uterine tumors (>4%), with “mutations” dominating the pattern. The “amplification” type of CNA was the primary type in the ovarian epithelial tumor cases, which show an alteration frequency of approximately 2%, It is worth noting that all miscellaneous neuroepithelial tumor, pheochromo- cytoma, and prostate cancer cases with genetic alteration (~2% frequency) had copy number deletion of MTF2 (Fig 4A). The types, loci and number of cases of MTF2 genetic alterations are
A
| Cancer Code | pvalue | Hazard Ratio(95% CI) | |
|---|---|---|---|
| TCGA-LGG (N=474) | 2.2e-6 . I- -I | 2.12(1.55,2.89) | |
| TCGA-GBMLGG (N=619) | 3.1e-5 | ¥ · I -I | 1.56(1.27,1.92) |
| TCGA-ACC (N=77) | 6.6e-5 I I | 2.87(1.71,4.80) | |
| TCGA-LIHC (N=341) | 1.4e-4 | . 1. -1 | 1.44(1.19,1.73) |
| TCGA-KIPAN (N=855) | 2.5e-3 I- -I | 1.30(1.10,1.54) | |
| TCGA-SARC(N=254) | 0.07 | 1 HI | 1.27(0.98,1.64) |
| TCGA-KICH(N=64) | 0.09 | I - | I 2.37(0.87,6.47) |
| TCGA-PAAD(N=172) | 0.16 | F 1 | 1.26(0.91,1.74) |
| TCGA-PRAD(N=492) | 0.22 | I · 1 | 1.86(0.69,5.05) |
| TCGA-UVM(N=74) | 0.25 | + 1 | 1.23(0.86,1.76) |
| TCGA-MESO(N=84) | 0.28 | I., -l | 1.23(0.85,1.79) |
| TCGA-ESCA(N=175) | 0.34 | I- -I | 1.20(0.83,1.74) |
| TCGA-KIRP(N=276) | 0.40 | -I | 1.18(0.80,1.75) |
| TCGA-THCA(N=501) | 0.46 | 1 1 | 1.37(0.59,3.18) |
| TCGA-SKCM-P(N=97) | 0.56 | 1- -- I | 1.09(0.81,1.46) |
| TCGA-PCPG(N=170) | 0.63 | I - -1 · | 1.37(0.38,4.99) |
| TCGA-CESC(N=273) | 0.69 | + 土 | 1.07(0.76,1.51) |
| TCGA-KIRC(N=515) | 0.89 | F 1 | 1.02(0.81,1.27) |
| TCGA-LUSC(N=468) | 0.89 | F 1 | 1.02(0.81,1.28) |
| TCGA-STES(N=547) | 0.90 | 1. I | 1.01(0.83,1.25) |
| TCGA-UCEC(N=166) | 0.91 | I 1 | 1.02(0.68,1.53) |
| TCGA-GBM(N=144) | 0.99 | I- -I | 1.00(0.76,1.32) |
| TCGA-THYM(N=117) | 0.03 | I l' | 0.54(0.30,0.96) |
| TCGA-SKCM(N=444) | 0.04 | + 1 | 0.87(0.76,1.00) |
| TCGA-OV(N=407) | 0.04 | F i | 0.87(0.76,0.99) |
| TCGA-SKCM-M(N=347) | 0.13 | F H | 0.89(0.76,1.04) |
| TCGA-COADREAD(N=368) | 0.16 | F | 0.74(0.48,1.13) |
| TCGA-READ(N=90) | 0.18 | + 1 | 0.56(0.24,1.29) |
| TCGA-LAML(N=144) | 0.37 | I- :- I | 0.87(0.64,1.18) |
| TCGA-UCS(N=55) | 0.40 | + 1 . | 0.76(0.40,1.45) |
| TCGA-COAD(N=278) | 0.41 | I I | 0.81(0.48,1.35) |
| TCGA-STAD(N=372) | 0.43 | F 1 | 0.90(0.70,1.16) |
| TCGA-HNSC(N=509) | 0.44 | I- -I | 0.93(0.78,1.11) |
| TCGA-CHOL(N=33) | 0.45 | I -1 | 0.80(0.44,1.44) |
| TCGA-BRCA(N=1044) | 0.78 | F ト | 0.97(0.76,1.22) |
| TCGA-LUAD(N=490) | 0.85 | I ·I | 0.98(0.79,1.21) |
| TCGA-TGCT(N=128) | 0.87 I | I | 0.86(0.16,4.76) |
| TCGA-BLCA(N=398) | 0.95 | I- -I | 0.99(0.81,1.22) |
| TCGA-DLBC(N=44) | 0.95 | I -1 | 0.98(0.45.2.12) |
T
Y
T
-25-20 -15 -1.0-0.5 0.0 0.5 1.0 1.5 20 25 log2(Hazard Ratio(95% CI))
B
C
D
Survival probability
1.0
ENSG00000143033(MTF2)
Survival probability
1.0
ENSG00000143033(MTF2)
0.8
L
0.8
L
Survival probability
1.0
ENSG00000143033(MTF2)
L
H
H
0.8
H
0.5
0.5
0.5
0.3
p=0.17
0.3
p=0.02
0.3
p=2.80-6
0.0
HR=1.21,95C1%(0.92,1.58)
0.0
HR=0.74,95C1%(0.58,0.96)
0.0
HR=232,95C1%(1.62,3.33)
Number at risk:
Number at risk
Number at risk
L
275
59
11
2
1
L
144
42
6
1
1
L
210
75
3
6
1
H
169
38
10
4
1
H
263
97
19
3
1
H
131
32
7
1
1
0
2,813
5,626
Õ
918
1,836
2,754
3,672
Overall survival
8,439
11,252
0
1,370
2,740
4,110
5,480
Overall survival
E
SKCM
F
Overall survival
OV
G
LIHC
Survival probabili
1.0
ENSG00000143033(MTF2)
1.0
ENSG00000143033(MTF2)
1.0
ENSG00000143033(MTF2)
0.8
Survival probability
Survival probability
H
0.8
L
0.8
L
H
≥0.5
0.5
0.5
0.3
p=1.2e-5
0.0
HR=7.25,95C1%(2.65,19.84)
0.3
0.3
p=1.50-3
p=1.004
Number at risk
0.0
HR=1.65,95C1%(1.21,2.26)
0.0
HR=2.29,96C1%(1.49,3.52)
L
35
24
12
3
1
Number at risk
Number at risk
H
$2
17
2
2
L
251
110
35
2
1
L 185
30
9
2
1
H
604
225
39
2
3
H
289
43
11
2
1
0
1.168
2,336
3,504
4,672
Overall survival
Ở
1.481
2,962
4.443
5,924
Ő
1,605
3,210
4,815
6,420
Overall survival
Overall survival
ACC
KIPAN
LGG
A
10
Group
..
Expression
Stage II
5
Stage I
0
Stage III
LUAD(Stagel=274,ll=122,l||=83,IV=26) COAD(Stage l=44,ll=110,Ill=82,IV=40)
COADREAD(Stage |=56,ll=134,lll=115,IV=53)
BRCA(Stage |=182,Il=617,|||=248,IV=20)
ESCA(Stagel=18,ll=80,Ill=61,IV=16)
KIPAN(Stage |=464,ll=107,|||=189,IV=103)
STAD(Stage |=58,ll=121,|||=169,IV=41)
KIRC(Stage |=266,11=57,|||=123,IV=81)
THYM(Stage|=36,ll=61,|||=14,IV=6)’
LIHC(Stage|=169,l|=86,|||=85,IV=5)
THCA(Stage|=283,11=52,|||=112,IV=55)
MESO(Stage |=10,Il=16,|||=45,IV=16)’
PAAD(Stage |=21,ll=147,Il|=3,IV=4)
SKCM(Stage ||=66,|||=26,IV=3) ACC(Stage |=9,Il=36,|||=15,IV=15)
Stage IV
B
10
Group
G3
5
G2
Expression
G1
0
G4
-5
-10
GBMLGG(G2=247,G3=260)
LGG(G2=247,G3=260)
CESC(G1=18,G2=135,G3=118)’
ESCA(G1=18,G2=74,G3=49)’
STES(G1=30,G2=222,G3=294)’
KIPAN(G1=14,G2=228,G3=206,G4=74)’
STAD(G1=12,G2=148,G3=245)’
UCEC(G1=14,G2=21,G3=141)”
HNSC(G1=61,G2=304,G3=124,G4=7)”
KIRC(G1=14,G2=228,G3=206,G4=74)”
LIHC(G1=55,G2=177,G3=121,G4=11)”
PAAD(G1=31,G2=95,G3=48)’
OV(G2=47,G3=360)
CHOL(G2=15,G3=18)’
C
PAAD(N=178)
SampleSize
D
ACC(N=77)
KIRP(N=285)
200
4.0
POPG(N=177)
400
BRCA(N=1090)
LGG(N=508)
LUAD(N=494)
600
SARC(N=258)
800
THCA(N=504)
3.5
KIRC(N=530)
- 1,000
BRCA(N=1090)
LGG(N=508)
ESCA(N=181)
pValue
PAAD(N=178)
LUAD(N=494)
BLCA(N=407)
0.0
3.0
LAML(N=173)
0.2
THCA(N=504)
GBM(N=152)
LIHC(N=368)
0.4
KIRP(N=285)
KIRC(N=530)
DLBC(N=47)
THYM(N=118)
0.6
-log10(pValue)
2.5
GBMLGG(N=660)
0.8
GBMLGG(N=660)
CHOL(N=36)
COAD(N=286)
BLCA(N=407)
KICH(N=66)
HNSC(N=517)
1.0
$2.0
THYM(N=118)
LUSC(N=489)
PRAD(N=496)
SARC(N=258)
COADREAD(N=377)
CESC(N=304)
UCEC(N=177)
KIPAN(N=881)
PCPG(N=17)
COAD(N=286)
UVM(N+79)
STES(N=500)
1.5
UCS(N=57)
OV(N=419)
ACC(N=77)
HNSC(N=517)
TGCT(N=132)
STAD(N=409)
SKCM(N=102)
LIHC(N=368)
PRAD(N=495)
UVM(N=79)
1.0
CESC(N=304)
LAML(N=173)
ESCA(N=181)
KIPAN(N=881)
UCS(N=57)
LUSC(N=489)
GBM(N=152)
STES(N=590)
MESO(N=87)
READ(N=91)
0.5
COADREAD(N=377)
OV(N=419)
MESO(N=87)
READ(N=91)
STAD(N=409)
DLBC(N=47)
KICH(N=66)
UCEC(N=177)
CHOL(N=36)
0.0
TGCT(N=132)
SKCM(N=102)
-0.2
-0.1
0.0
0.1
0.2
-0.20
-0.15
-0.10
-0.05
0.00
0.05
0.10
0.15
Correlation coefficient(pearson)
Correlation coefficient(pearson)
A
5%
Alteration Frequency
Mutation
Structural Variant
4%-
Amplification
Deep Deletion
3%
2%-
1%-
Structural variant data Mutation data CNA data
Endometrial Cancer
Miscellaneous Neuroepithelial Tumor
Sarcoma
Bladder Cancer
Ovarian Epithelial Tumor
Melanoma
Esophagogastric Cancer
Adrenocortical Carcinoma
Colorectal Cancer
Pheochromocytoma
Breast Cancer
Non-Small Cell Lung Cancer
Seminoma
Cervical Cancer
Non-Seminomatous Germ Cell Tumor
Prostate Cancer
Pancreatic Cancer
Hepatobiliary Cancer
Leukemia
Thyroid Cancer
Head and Neck Cancer
Glioblastoma
Renal Clear Cell Carcinoma
Glioma
Cholangiocarcinoma
Mature B-Cell Neoplasms
Renal Non-Clear Cell Carcinoma
Pleural Mesothelioma
Thymic Epithelial Tumor
Ocular Melanoma
B
GBM(N=149,0.7%)
Missense_Mutation
GBMLGG(N-649,0.2%)
Frame Shift Ins
CESC(N-286,1.0%)
In_Frame_Ins
LUAD(N-508,0.8%)
In_Frame_Del
COAD(N-282,2.1%)
Frame_Shift_Del
COADREAD(N-372.2.4%)
Nonsense_Mutation
BRCA(N-980,0.3%)
Splice_Site
ESCA(N-180.0.6%)
3.0
STES(N=589.1.9%)
SARC(N-234,0,4%)
STAD(N=409,2.4%)
2.5
UCEC(N-175,2.9%)
HNSC(N-498,0.4%)
2.0
LUSC(N=485,0.8%)
LIHC(N-356,0.3%)
THCA(N-487,0.2%)
1.5
READ(N-90.3.3%)
PAAD(N-168.0.6%)
OV(N-303,0.7%)
1.0
TGCT(N-143.1.4%)
BLCA(N=407.1.2%)
593aa
PHD
TUDOR
PHD
PHD
Mtf2_C
Mtf2_C
C
100%
Logrank Test P-Value: 0.0476
D
100%
Disease Free rate
90%
80%
Disease-specific rate
90%
Logrank Test P-Value: 0.163
80%
70%
70%
60%
60%
50%
50%
40%
40%
30%
30%
20%
Disease Free
20%
Disease-specific
10%
Altered group
10%
0%
Unaltered group
· Altered group
0
20 40 60 80 100120140160180200220 Disease Free (Months)
0%
20
Unaltered group
E
0
40
60
80
100
201401
160180200220
F
Months of disease-specific survival
Probability of Overall Survival rate
100%
100%
90%
Logrank Test P-Value: 0.0561
Progression Free rate
90%
80%
80%
Logrank Test P-Value: 0.0381
70%
70%
60%
60%
50%
50%
40%
40%
30%
30%
20%
Overall
20%
Altered group
Progression Free
10%
0%
Unaltered group
10%
Altered group
Unaltered group
0
20
40
60
80
100 120 140 160 180 200 220
0%
0
20
40
60
80
Overall Survival (Months)
100 120 140 160 180 200 220
Progress Free Survival (Months)
further described in Figure 4B. We found major missense mutations at 2 MTF2 sites. Among GBM and LGG cases, only missense mutations were present at MTF2 sites (Fig 4B). To understand if there is a relationship between certain genetic alterations in MTF2 and patient outcomes, we investigated UCEC tumors using the cBioPortal tool. We found that patients with MTF2-altered had a favorable prognosis for progression-free survival and disease-free survival compared to patients without MTF2 alterations, but there was no difference between OS and disease-specific survival (P < 0.05, Fig 4C-G)
RNA-Modified Gene Analysis
Due to the lack of studies on the role of MTF2 in tumor function, this study sought to begin with functional analysis of MTF2-associated genes. Chemical modification is an effective way to regulate the function of large molecules such as DNA, RNA and proteins. These macromolecules require post-synthesis and covalent modifications to function in organisms. Sequencing revealed that RNA modification plays a key role in selective gene expression.24 Pearson correlation anal- ysis of MTF2 and marker genes revealed that MTF2 is associated with marker genes in most tumors. In addition, we found positive correlations between LUSC, STAD, STES, ACC, OV tumors in all m1A-, m5C-, and m6A-related genes by the expression of MTF2 genes and 44 marker genes of 3 types of RNA modifications (m1A,10 m5C,13 m6A21) genes in pan-cancer, as shown in Figure 5 with a high correlation, especially higher in COADREAD and COAD tumors (P < 0.05, Figure 5. (See Supplementary Table 17 for statistical details).
Methylation Analysis
DNA methylation directly affects cancer initiation and progression,25 so we analyzed MTF2 methylation levels in both normal and primary tumor tissues from the CPTAC database. Methy- lation levels were significantly higher in CHOL, COAD, ESCA, HNSC, and UCEC than in normal tissues (P < 0.05, Fig 6D-H). However, BLCA, BRCA and CESC methylation levels were not signif- icantly associated with corresponding normal tissues (Fig 6A-C).
Immunocheckpoint Analysis
In order to elucidate the relationship between MTF2 expression and specific immune cell types in cancer. We selected 60 immune checkpoint genes for analysis, including inhibitory (n = 24) and stimulatory (n = 36) genes. Interestingly, the results of immune checkpoint analysis showed that MTF2 expression was positively correlated with most immune checkpoint genes. In this respect, more than 50 immune checkpoint genes were positively correlated with MTF2 ex- pression in KIRC, KIPAN, PRAD, LIHC, COADREAD and PAAD. In addition, in CHOL, the expression of MTF was least correlated with the immune checkpoint genes (P < 0.05, Fig 7). (See Supple- mentary Table 18 for statistical details).
Immunocyte Analysis
Further to our analysis of immune cells, we observed a significant and mostly positive as- sociation between MTF2 expression and immune infiltration in 35 cancer types, such as ACC, BRCA, CESC, COAD, COADREAD, DLBC (lymphoid neoplasm diffuse large B-cell lymphoma), ESCA, GBMLGG, HNSC, KICH, KIPAN, KIRC, KIRP, LGG, LIHC, LUAD, LUSC, mesothelioma (MESO), OV, PAAD, PCPG, PRAD, rectum adenocarcinoma (READ), sarcoma (SARC), skin cutaneous melanoma- metastasis (SKCM-M), skin cutaneous melanoma-primary (SKCM-P), skin cutaneous melanoma
Modification
Type
TRMT61A
correlation coefficient
TRMT6
TRMTIOC
TRMT61B
-1.0-0.5 0.0 0.5 1.0
YTHDFI
pValue
YTIIDC1
YTHDF2
YTHDF3
0.0
0.5
1.0
ALKBIII
Modification:
ALKBH3
NOP2
mlA
DNMTI
m5C
NSUN6
m6A
NSUN3
Type:
TRDMT1
writer
NSUN2
reader
DNMT3B
eraser
NSUN4
DNMT3A
NSUNS
NSUN7
TET2
ALYREF
KIAA1429
RBM15
RBM15B
WTAP
ZC31113
METTI.3
METTL14
CBLLI
ALKBH5
FTO
YTHDFI
HNRNPC
ELAVLI
FMRI
YTHDC2
YTHDF3
YTHDC1
YTHDF2
HNRNPA2B1
IGF2BP1
TGCT(N=148)
LUAD(N=513)
LUSC(N=498) ESCA(N=181)
RPPRC
STAD(N=414)
STES(N=595
MESO(N=87
BLCA(N=407)
BRCA(N=1092)
READ(N=92)
GBMLGG(N=662)
LGG(N=509)
LAML(N=173)
CHOL(N=36)
UCS(N=57
PCPG(N=177)
COAD(N=288)
COADREAD(N=380)
DLBC(N=47)
THYM(N=119)
KICH(N=66)
KIPAN(N=884)
KIRC(N=530)
CESC(N=304) HNSC(N)
SARC(N=258)
PRAD(N=495)
UCEC(N=180)
KIRP(N=288)
THCA(N=504)
LIHC(N=369)
PAAD(N=178)
SKCM(N=102)
UVM(N=79)
GBM(N=153)
ACC(N=17
OV(N=419)
(SKCM), STAD, STES, testicular germ cell tumors (TGCT), thyroid carcinoma (THCA), THYM, UCEC, uterine carcinosarcoma (UCS) and uveal melanoma (UVM). MTF2 expression was associated with B cell, CD4 T cell, CD8 T cell, neutrophil, macrophage, and DC in 18, 25, 24, 29, 19, and 22 tu- mors, respectively (P < 0.05, Fig 8). (See Supplementary Table 19 for statistical details).
Immune Infiltration Analysis
To further evaluate the role of MTF2 in the tumor immune microenvironment, we analyzed the relationship between MTF2 expression and immune infusion score in tumors using ESTI- MATE. We observed a significant association between MTF2 expression and purity in 21 tumors, including 6 positive associations such as LGG, COAD, COADREAD, KIPAN, KIRC, PAAD, and 15 negative associations such as GBM, CESC, LUAD, LAML, STES, SARC, KIRP, UCEC, LUSC, THCA, OV, TGCT, SKCM-P, BLCA, ACC (P < 0.05, Fig 9A) . Among 22 cancer types, MTF2 expression was sig- nificantly associated with immune cell infiltration, including 6 positive associations such as LGG, COAD, COADREAD, KIPAN, KIRC, PAAD, and 16 negative associations such as GBM, CESC, LUAD, LAML, STES, SARC, KIRP, UCEC, LUSC, THCA, OV, TGCT, SKCM-P, UCS, BLCA, ACC (P < 0.05, Fig 9B) . MTF2 expression was significantly associated with stromal cell infiltration in 23 cancer types,
A
B
Promoter methylation of MTF2 in BLCA
Promoter methylation of MTF2 in BRCA
0.07
0.08
P=9.43E-01
P=9.35E-01
0.06
0.07
Bate value
0.05
Bate value
0.06
0.05
0.04
0.04
0.03
0.03
0.02
Normal n=21
Primary tumor n=418
0.02
TCGA Samples
Normal n=97
TCGA Samples
Primary tumor n=793
C
D
0.09
Promoter methylation of MTF2 in CESC
Promoter methylation of MTF2 in CHOL
0.1
0.08
P=2.07E-01
P=7.20E-03
0.08
Bate value
0.07
Bate value
0.06
0.06
0.05
0.04
0.04
0.03
0.02
Normal n=3
Primary tumor n=307
0.02
Normal n=9
Primary tumor n=36
TCGA Samples
TCGA Samples
E
F
0.08
Promoter methylation of MTF2 in COAD
0.08
Promoter methylation of MTF2 in ESCA
0.07
0.07
P=1.64E-12
P=8.56E-08
0.06
Bate value
Bate value
0.06
0.05
0.05
0.04
0.04
0.03
0.03
0.02
0.02
Normal n=37
Primary tumor n=313
0.01
Normal n=16
Primary tumor n=18
TCGA Samples
TCGA Samples
G
H
0.09
Promoter methylation of MTF2 in HNSC
0.07
Promoter methylation of MTF2 in UCEC P=7.74E-03
0.08
P=9.18E-12
0.06
Bate value
0.07
Bate value
0.06
0.05
0.05
0.04
0.04
0.03
0.03
0.02
Normal n=50
Primary tumor n=528
0.02
Normal n=46
Primary tumor n=438
TCGA Samples
TCGA Samples
Type
VEGFB
LAG3
IDO1
KIR2DL 1
KIR2DL3
PDCD1
SLAMF7
CTLA4
TIGIT
CD274
IL 10
BTLA
HAVCR2
IL4
ARG1
IL 13
VTCN1
VEGFA
TGFB1
C10orf54
ADORA2A
IL 12A
CD276
EDNRB
HMGB1
ENTPD1
TLR4
BTN3A1
TNFSF4
IFNA1
IFNA2
TNFRSF14
TNFRSF18
TNFRSF4
CD27
GZMA
CCL5
TNFSF9
CD70
IL2
IL 1A
IL 1B
CX3CL 1
CD40
TNF
ICOSLG
SELP
BTN3A2
CD28
ICOS
IFNG
CXCL 10
CXCL9
IL2RA
CD80
ICAM1
TNFRSF9
ITGB2
CD40L G
OV(N=419)
PAAD(N=178)
KIPAN(N=884)
KIRC(N=530)
PRAD(N=495)
READ(N=92)
KICH(N=66)
LIHC(N=369
DLBC(N=47)
UVM(N=79)
THYM(N=119)
LAML(N=173)
CHOL(N=36)
SKCM(N=102) UCEC(N=180)
PRF1
GBML GG(N=662)
LGG(N=509)
ESCA(N=181)
COAD(N=288)
COADREAD(N=380)
STES(N=595 STAD(N=414)
MESO(N=87)
BRCA(N=1092)
HNSC(N=518
TGCT(N=148)
UCS(N=57)
CESC(N=304)
BLCA(N=407)
SARC(N=258) GBM(N=153)
LUAD(N=513)
LUSC(N=498)
ACCIN=77
PCPG(N=177)
KIRP(N=288)
THCA(N=504)
correlation coefficient
-1.0-0.5 0.0 0.5 1.0 pValue
0.0
0.5
1.0
Type:
Inhibitory
Stimulaotry
correlation coefficient
T
T
T
-1.0-0.5 0
0.0
0.5
1.0
pValue
0.0
0.5
1.0
0.46
0.18
0.23
0.43
0.25
0.28
0.43
0.24
0.58
0.54
0.38
0.47
0.38
0.13
0.21
0.33
0.18
0.24
0.63
0.46
0.62
0.27
0.64
**
0.32
0.60
0.22
0.51
0.40
0.47
0.30
0.32
0.14
0.45
0.32
0.37
0.35
0.34
0.48
0.52
0.26
0.50
0.19
0.51
0.39
0.51
0.20
0.12
0.18
0.23
0.32
0.20
0.23
0.40
0.23
0.29
0.38
*
**
0.27
0.30
0.49
0.25
0.37
0.12
0.43
0.09
0.36
0.20
0.28
**
0.18
0.26
0.23
0.29
0.18
0.25
0.41
0.42
0.45
0.48
0.19
0.28
0.25
0.30
0.17
0.26
**
**
-0.19
**
0.13
0.25
0.29
0.17
0.12
0.14
0.11
0.33
0.11
**
**
*
*
0.26
0.24
0.39
0.28
0.33
*
*
**
**
0.22
0.11
0.17
*
0.13
0.32
*
0.18
0.12
0.21
0.18
*
0.16
-0.20
0.27
*
**
0.26
0.28
0.30
0.11
*
0.16
0.12
0.23
0.16
**
-0.24
0.12
-0.16
*
*
0.09
0.24
0.09
*
*
0.30
**
0.28
0.31
**
**
0.21
0.23
-0.19
0.21
*
*
0.56
0.23
*
0.24
0.23
**
**
-0.28
0.34
*
*
B cell
T cell CD4
T cell CD8
Neutrophil
Macrophage
DC
TCGA-LGG(N=504)
TCGA-PRAD(N=495)
TCGA-GBMLGG(N=656)
TCGA-THYM(N=118)
TCGA-KIRC(N=528)
TCGA-LIHC(N=363)
TCGA-PAAD(N=177)
TCGA-THCA(N=503)
TCGA-BRCA(N=1077)
TCGA-MESO(N=85)
TCGA-PCPG(N=177)
TCGA-KIPAN(N=878)
TCGA-COADREAD(N=373)
TCGA-KICH(N=65)
TCGA-COAD(N=282)
TCGA-CESC(N=291)
TCGA-HNSC(N=517)
TCGA-SKCM(N=452)
TCGA-READ(N=91)
TCGA-LUSC(N=491)
TCGA-SKCM-M(N=351)
TCGA-STAD(N=388)
TCGA-UCEC(N=178)
TCGA-OV(N=417)
TCGA-GBM(N=152)
TCGA-STES(N=569)
TCGA-SARC(N=258)
TCGA-LUAD(N=500)
TCGA-BLCA(N=405)
TCGA-UVM(N=79)
TCGA-SKCM-P(N=101)
TCGA-KIRP(N=285)
TCGA-TGCT(N=132)
TCGA-DLBC(N=46)
TCGA-CHOL(N=36)
TCGA-ACC(N=77)
TCGA-ESCA(N=181)
TCGA-UCS(N=56)
including 8 with significant positive associations such as LGG, COAD, COADREAD, KIPAN, KIRC, MESO, PAAD, PCPG, and 15 negative associations such as GBM, CESC, LUAD, LAML, STES, SARC, KIRP, STAD, UCEC, LUSC, OV, TGCT, SKCM-P, BLCA, ACC (P < 0.05, Figure 9C) . (See supplementary table 20 for statistical details).
A
OTCGA-KIPAN(N=878): p=1.2e-17 r=0.28
6,000
OTCGA-BLCA(N=405): p=3.3c-12 r =- 0.34
OTCGA-KIRC(N=528): p=5.8e-9 r=0.25
OTCGA-SARC(N=258): p=1.1c-7r =- 0.32
OTCGA-TGCT(N=132): p=2.2c-7 r =- 0.43
4,000 -
OTCGA-GBM(N=152): p=1.le-6r =- 0.38
TCGA-KIRP(N=285): p=1.3c-6 r =- 0.28
OTCGASESC(N=291): p=1.4e-5 r =- 0.25
ESTIMATEScore
2,000
OTCGA-OV(N 417) 1.76-5 r -0.21
OTCGA-UCEC(N=178): p 3.88-55-0.30
OTCGA-LUSC(N=491): p=5.2c-5 r =- 0.18
OTCGA-STES(N=569): p=1.3c-4 r =- 0.16
0 -
TCGA-LAML(N=149): p=2.0c-4r =- 0.30
OTCGA-PAAD(N=177): p=4.0c-4 r=0.26
OTCGA-SKCM-P(N=101): p=8.6c-4 r =- 0.33
-2,000 -
OTCGA-THCA(N=503): p=1.0e-3 r =- 0.15
TCGA-LGG(N=504): p=1.3e-3 r=0.14
OTCGA-LUAD(N=500): p=2.1e-3 r =- 0.14
OTCGA-COAD(N=282): p=3.6e-3 r=0.17
-4,000 -
OTCGA-ACC(N=77): p=4.6e-3 r =- 0.32
OTCGA-COADREAD(N=373): p=9.3e-3 r=0.13
-5
0
5
B
MTF2 Expression
TCGA-KIPAN(N=878): p=1.0c-10 r=0.22
OTCGA-BLCA(N=405): p=6.7e-10 r =- 0.30
4,000
OTCGA-SARC(N=258): p=2.3e-8 r =- 0.34
OTCGA-GBM(N=152): p=4.8c-8 r =- 0.43
OTCGA-CESC(N=291): p=2.2c-7r =- 0.30
TCGA-KIRP(N=285): p=2.3e-7 r =- 0.30
OTCGA-UCEC(N=178): p=8.6c-7 r =- 0.36
STOGA-OV(N=417): p=2.1c-6 r =- 0.23
TCGA-THCAS=503): p=1.le-5 r =- 0.19
ImmuneScore
2,000
TCGA-LAML(N=149): p=1-5c-5 r =- 0.35
OTCGA-TGCT(N=132): p=4.6c-5 1-35
OTCGA-LUSC(N=491): p=8.9c-5 r =- 0.18
OTCGA-LGG(N=504): p=2.6e-4 r=0.16
OTCGA-KIRC(N=528): p=1.1c-3 r=0.14
OTCGA-LUAD(N=500): p=1.le-3 r =- 0.15
0
TCGA-PAAD(N=177): p=1.6c-3 r=0.24
OTCGA-ACC(N=77): p=1.8c-3 r =- 0.35
OTCGA-SKCM-P(N=101): p=3.3c-3 r =- 0.29
OTCGA-STES(N=569): p=6.4c-3 r =- 0.11
TCGA-COAD(N=282): p=6.8e-3 r=0.16
OTCGA-COADREAD(N=373): p=0.01 r=0.13
2,000
OTCGA-UCS(N=56): p=0.02 r =- 0.32
5
0
5
MTF2 Expression
C
TCGA-KIPAN(N=878): p=2.9e-23 r=0.33
OTCGA-KIRC(N=528): p=6.0e-15 r=0.33
2,000
TCGA-BL.CA(N=405): p 8.4e-12 r -0.33
OTCGA-STES(N=569): p=1.2e-5r — 0.18
OTCGA-TGCT(N=132): p=1.6e-5 r =- 0.37
OTCGA-SARC(N=258): p=9.0e-5 r-0.24
1,000
OTCGA-LUSC(N=491): p=2.2e-4r =- 0.17
TCGA-KIRP(N=285): p=2.4e-4r =- 0.22
OTCOM-PCRG(N=177): p=2.5e-4 r=0.27
StromalScore
TCGA-GBM(N=152): p=2.60-4r =- 0.29
0
TCGA-PAAD(N=177): p-3.8e-4 r 0.26
TCGA-SKCM-P(N=101): p=1.2e-3 r =- 0.32
TCGA-STAD(N=388): p=2.6e-3 r =- 0.15
-1,000
OTCGA-OV(N=417): p=2.9c-3 r =- 0.15
TCGA-COAD(N=282): p=5.3c-3r=0.17
TCGA-COADREAD(N=373): p=0.02 r=0.12
TCGA-LUAD(N=500): p=0.02 r =- 0.11
-2,000
TCGA-LAML(N=149): p=0.02 r =- 0.19
TCGA-LGG(N=504): p=0.02 r=0.10
TCGA-ACC(N=77): p=0.03 r =- 0.25
OTCGA-CESC(N=291): p=0.04 r =- 0.12
-3,000
OTCGA-MESO(N=85): p=0.04r=0.22
TCGA-UCEC(N=178): p=0.05 r =- 0.15
-5
0
5
MTF2 Expression
PPI Analysis
To investigate this profound effect of MTF2 on tumorigenesis and progression, we used the STRING tool to analyze MTF2 interacting proteins, and we screened the top 10 MTF2 binding proteins. PPI of these 10 proteins was also shown. (Fig 10A) . Next, we used the GEPIA2 tool to integrate all tumor expression data from TCGA and obtained MTF2-associated scatter plot. Results showed that MTF2 was positively associated with expression of EED, EZH2, SUZ12, RBBP4 RBBP7, CBX4, JARID2, AEBP2, C17orf96, and EZH1 in TCGA tumors. (P < 0.05) (Fig 10B-K) .
Cytology Western Blot Validates MTF2 Biological Function
To further validate the expression of MTF2-interacting genes and the role of MTF2 in can- cer, we transfected hMTF2, shMTF2, and vector lentivirus into U87, U25, and MDA-MB-231 cell lines to construct stable cell lines. The expression levels of MTF2, PPI reciprocal genes EZH2 and EMT-associated genes N-Cadherin and vimentin were then measured by Western blot. Results showed that MTF2, EZH2, N-Cadherin, and vimentin expression were significantly upregulated in U87, U251, and MDA-MB-231 cell lines compared to the vector group (P < 0.05, Fig 11A), while in both U251 and MDA-MB-231 cell lines, expression of these gene proteins was signifi- cantly downregulated in shMTF2 (P < 0.05, Fig 11B).
Cytological phenotype Analysis of MTF2 Biological Function
To further validate MTF2 migration and invasion of tumors, we performed cytological HE, panel clonogenic assays, wound healing assay, and transwell assays. HE results showed that U87 cells were fully formed and in good condition after transfection with vector and hMTF2 (Fig 12B). Plate clonogenic assays in U87 and MDA-MB-231 cell lines showed a significant in- crease in the number of clones in hMTF2 compared to the vector group. In the shMTF2 group, the reverse was true (P < 0.05, Fig 12A, D). U87 and MDA-MB-23 cell lines scratched signifi- cantly increased migration in the hMTF2 group compared to the Vector group, but not in the shMTF2 group (P < 0.05, Fig 12C, E). In the MDA-MB-23 cell line, the transwell showed signifi- cantly increased invasion in the hMTF2 group compared to the Vector group (P < 0.05, Fig 12F).
Discussion
MTF2 plays a key role in cell development.26 A large body of literature has shown that MTF2 is key to regulating involvement in ESC self-renewal and differentiation.27 Decreased MTF2 con- tent resulted in enhanced self-renewal properties and inefficient differentiation of the 3 em- bryos.28 In fruit fly embryos, PCL and PRC2 formed complexes and maximized their catalytic activity on polycomb target genes.29 It was also revealed that MTF2 is an important epigenetic regulator of Wnt signaling pathway during erythrogenesis.30 Recent studies have reported that MTF2-overexpressing leukemia cells are highly sensitive to chemotherapy-induced relapse and that their response to MDM2 inhibitors that overexpress or target signaling pathways is reduced in AML-derived xenograft mouse models.31 In recent years, many studies have linked MTF2 to tumors, including glioma and liver cancer.14,32 However, it is not known whether MTF2 is in- volved in other tumorigenesis or, more specifically, which tumorigenic malignancies it plays a role in. Therefore, we comprehensively analyzed MTF2 pan-cancer.
Using the TCGA and GTEx datasets, we show that MTF2 genes are highly expressed in a total of 22 cancers and low in 8, with LUAD inconsistently expressed in both databases, possibly due to differences in samples. The GBM and LIHC results were similar to previous studies,14,32 and
SUZ12
A
EZH1
C17orf96
RBBP4
B
-
RBBP7
p-value = 0
AEBP2
R = 0.57
.
CBX4
log2(SUZ12 TPM)
EED
៛
MTF2
2
EZH2
JARID2
0
0
1
2
3
4
5
6
7
log2(MTF2 TPM)
C
D
E
-
.
p-value = 0
A
p-value = 0
p-value = 0
>
R = 0.45
R = 0.49
..
R = 0.41
.
6
”
log2(EED TPM)
₼
log2(EZH2 TPM)
log2(EZH1 TPM)
5
4
+
.
·
”
0
0
2
2
2
-
-
-
0
0
.
0
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
log2(MTF2 TPM)
log2(MTF2 TPM)
log2(MTF2 TPM)
F
G
H
º
80
p-value = 4.2e-32
p-value = 0
R = 0.6
p-value = 0
R =- 0.12
.
R = 0.12
log2(PRELID1 TPM)
-
log2(JARID2 TPM)
.
log2(CBX4 TPM)
”
+
.
2
~
2
0
o
0
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
log2(MTF2 TPM)
G
log2(MTF2 TPM)
K
log2(MTF2 TPM)
0
p-value = 0
2
p-value = 0
8
p-value = 1.3e-14
R = 0.54
R = 0.11
R = 0.078
:
8
log2(RBBP4 TPM)
log2(RBBP7 TPM)
log2(C17orf96 TPM)
0
8
.
0
.
+
+
~
2
~
0
0
0
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
0
1
2
3
4
5
6
7
log2(MTF2 TPM)
log2(MTF2 TPM)
log2(MTF2 TPM)
A
Vector
hMTF2
Relative Protein Expression
1.5
EZH2
98kDa
**
Vector
1.0
GAPDH
37kDa
hMTF2
*
MTF2
67kDa
GAPDH
0.5
37kDa
N-Cadherin
120kDa
U87
GAPDH
37kDa
0.0
EZH2
MTF2
N-Cadherin Vimentin
Vimentin
55kDa
1.5
GAPDH
37kDa
Relative Protein Expression
**
EZH2
98kDa
1.0
Vector
**
*
hMTF2
GAPDH
37kDa
MTF2
67kDa
0.5
GAPDH
37kDa
U251
N-Cadherin
120kDa
0.0
EZH2
MTF2
N-Cadherin Vimentin
Vimentin
55kDa
Relative Protein Expression
1.5
GAPDH
37kDa
Vector
**
**
MTF2
67kDa
1.0
hMTF2
GAPDH
37kDa
MDA-MB-231
0.5
N-Cadherin
120kDa
Vimentin
55kDa
0.0
MTF2
N-Cadherin Vimentin
GAPDH
37kDa
B
Vector
shMTF2
Relative Protein Expression
1.5
EZH2
98kDa
1.0
Vector
a-Actin
43kDa
shMTF2
*
MTF2
67kDa
0.5
GAPDH
37kDa
U251
N-Cadherin
120kDa
0.0
EZH2
MTF2
N-Cadherin Vimentin
GAPDH
37kDa
Relative Protein Expression
1.5
Vimentin
55kDa
Vector
GAPDH
37kDa
1.0
shMTF2
MTF2
67kDa
MDA-MB-231
N-Cadherin
120kDa
0.5
**
**
Vimentin
55kDa
0.0
GAPDH
37kDa
MTF2
N-Cadherin Vimentin
A
Vector
Number of colonies
400
hMTF2
B
Vector
hMTF2
300
200
100
0
Vector
hMTF2
C
Oh
24h
48h
72h
Wound healing percent(%)
1.5
U87
1.0
hMTF2
0.5
Vector
0.0
Vector hMTF2
D
Vector
hMTF2
shMTF2
300
Vector
Number of colonies
hMTF2
shMTF2
200
100
0
Vector
hMTF2
shMTF2
E
Vector
hMTF2
shMTF2
1.5
Wound healing percent(%)
0h
1.0
0.5
24h
0.0
Vector
hMTF2
shMTF2
MDA-MB-231
F
Vector
hMTF2
shMTF2
Number of invaded cells in 24h
400
300
200
100
0
Vector
hMTF2
shMTF2
we demonstrated this trend in LIHC at the protein level. In addition, MTF2 expression is upreg- ulated in PRAD cells,33 which contradicts our current results, possibly because the more diverse versions analyzed in our study were derived from in situ tumors rather than metastases, and previous studies included more highly proliferative cancer studies focusing on metastasis. Inter- estingly, when combined with survival analysis, we found poor prognosis for high expression in GBM, LGG, KIPAN, LIHC, ACC. Similarly, MTF2 expression was previously reported to be associ- ated with shorter survival in GBM, LIHC, multiple myeloma, RBs and PRAD patients.14,15,17,32,33 Overexpression of MTF2 sensitizes AML to chemotherapeutic agents, MTF2 deficiency predicts refractory AML at diagnosis. MTF2 represses MDM2 in hematopoietic cells and its loss in AML results in chemoresistance. Inhibiting p53 degradation by overexpressing MTF2 in vitro or by using MDM2 inhibitors in vivo sensitizes MTF2-deficient refractory AML cells to a standard induction-chemotherapy regimen.16 These results suggest that MTF2 may be a potential prog- nostic biomarker, but in vitro or in vivo validation assays are lacking. Therefore, the role of MTF2 in different cancer types remains to be further investigated.
Furthermore, we found that MTF2 expression is associated with age in certain types of cancer. In GBM LGG, LGG, LUAD, BRCA, ARC, KIRP, KIRC, THCA, PAAD, BLCA patients, MTF2 expression was negatively correlated with age. These results may have important implications for guiding treatment options for patients of different age groups. Our study also shows that MTF2 expres- sion is associated with tumor staging of a few cancers, such as BRCA, KIPAN, THYM, LIHC, ACC. We found MTF2 expression in both early and late stages of these cancers, but early MTF2 ex- pression was generally higher than late stages. Cancer often develops because of genetic changes. Therefore, we further analyzed the genetic alterations in MTF2. We have implications for future studies of the role of MTF2 mutations in cancer. MTF2 as an important epigenetic regulatory gene.9 Epigenetics is a stable inheritance of gene expression or functional alterations by regu- lating genome-environment interactions without altering basic DNA sequences, including DNA methylation, histone modification, chromatin remodeling, and RNA modification.34,35,38 Epige- netic abnormalities are considered to be one of the most important oncogenic mechanisms, so we further performed RNA modification analysis to analyze the correlation between MTF2 ex- pression and marker genes. Further, to determine if MTF2 expression plays a role in DNA methy- lation, we analyzed MTF2 expression at the DNA methylation level in association with MTF2 ex- pression. The results showed that methylation levels were significantly higher in CHOL, COAD, ESCA, HNSC, and UCEC than in normal tissues. Although the sample size of methylation stud- ies is insufficient here, combining multiple tumor data analyses suggests that MTF2 methylation levels may be a biomarker of prognosis in cancer patients.
Tumors are not only composed of malignant cells but also embedded in complex interacting microenvironments.37 Immunotherapy has become a new pillar of cancer treatment in recent years. The tumor immune microenvironment is an important component of the tumor microen- vironment, and the mechanisms by which tumor cells work with the immune microenviron- ment are important for the selection of key molecules for tumor markers and potential drug tar- gets.36,39 Tumor immunotherapy has dramatically changed the paradigm of cancer patient man- agement. Studies have shown that immune checkpoint blockade therapy improves survival in pa- tients with advanced melanoma, non-small cell lung cancer (NSCLC) and other cancers. In addi- tion, tumor-infiltrating lymphocytes (TILs), tumor-associated macrophages and tumor-infiltrating neutrophils (TINs), play an important role in tumor immunity.41,40 These findings confirm the importance of tumor immune infiltration in tumor progression, so analysis of the role of PCL2 in the immune microenvironment cannot be ignored. We found that PCL2 expression was pos- itively correlated with immune checkpoints in most tumors. Further, we analyzed immune cell infiltration scores (B cell, T cell CD4, T cell CD8, Neutrophil, Macrophage, DC), and ultimately we observed a significant correlation between expression of this gene and immune infiltration in 33 cancer types. Finally identifying significant correlation between stromal, immune, and ESTIMATE scores, we observed significant positive correlation between MTF2 gene expression and stromal cell infiltration in 8 cancer types, such as LGG, COAD, COADREAD, KIPAN, KIRC, MESO, PAAD, and PCPG. MTF2 gene expression was significantly positively associated with stromal cell infil- tration in six cancer types, such as LGG, COAD, COADREAD, KIPAN, KIRC, and PAAD. MTF2 gene
expression was significantly positively associated with tumor cell infiltration in 6 cancer types, including LGG, COAD, COADREAD, KIPAN, KIRC, and PAAD. We show that GABRD genes are in- volved in tumor immune microenvironment expression. Immune checkpoint genes have been reported to directly affect immune cell function.47 During tumorigenesis, immune escape check- points are activated by tumorigenesis to avoid attack, which leads to tumor aggressiveness.46 Therefore, analyzing the association of MTF2 expression with immune cells could provide new targets for studying tumor immune inhibitors. These results suggest that MTF2 plays an immune role in a wide range of tumors and may be a target for cancer therapy.
In addition, we analyzed MTF2-interacting gene networks using the STRING tool, and then we used the GEPIA2 tool to analyze the association of these genes with MTF2 expression. Re- sults showed that MTF2 expression was positively correlated with 10 genes. Many genes play different roles in different cancers.42-45 EZH2 is a histone methyltransferase (HMT) that cat- alyzes H3K27me2/3. PRC2-EZH2 regulates H3K27me2/3 levels in cells through its EZH2-mediated methyltransferase activity.49 Studies have found that genome-wide recruitment of the PRC2 cat- alytic subunit EZH2 is abrogated in Mtf2 knockout cells, resulting in greatly reduced H3K27me3 deposition.11 MTF2 can also interact with PRC 2 via EZH2.48 In addition, MTF2, as an important coenzyme of PRC2, affects the expression of core proteins EZH2 and changes histone (H3K27, H3K9, and H3K4) methylation. The effects of EZH2 can be enhanced by increasing MTF2 expres- sion, and this protein interaction is involved in changes in histone methylation.32
Studies have found that genome-wide recruitment of the PRC2 catalytic subunit EZH2 is ab- rogated in Mtf2 knockout cells, resulting in greatly reduced H3K27me3 deposition. MTF2 can also interact with PRC 2 via EZH 2. In addition, MTF 2, as an important coenzyme of PRC 2, affects the expression of core proteins EZH 2 and EED and changes histone (H3K27, H3K9 and H3K4) methylation. The effects of EZH 2 can be enhanced by increasing MTF 2 expression, and this protein interaction is involved in changes in histone methylation. Further validated by cy- tology, we constructed stable overexpressing hMTF2/ hMTF2 and Vector U87, U251, and MDA- MB-231 cell lines, and measured MTF2, EZH2, and EMT-associated protein expression levels via Western blot. Results showed that MTF2 and EZH2 expression were significantly increased in hMTF2 compared to the Vector group. We found that zeste homologous 2 (EZH2) enhancers are multipotent in cancer and immune cells, and that EZH2 plays a role in normal biology of many cell types, including immune cells.50 Dysfunctional EZH2 is associated with the develop- ment of multiple cancer types in mice and humans.51 The results also suggest that MTF2 may be a driver of malignant glioma progression. ShMTF2 does the opposite. These results suggest that MTF2 promotes GBM and BRCA migration invasion. To further validate the migration and invasion of MTF2 to tumors, we performed in vitro validation using cytological HE, panel clonal assays, scratch assays, and transwell phenotype assays. HE results showed that U87 cells were fully formed and in good condition after transfection with Vector and hMTF2. Results showed a significant increase in clones, migration and invasion in the hMTF2 group compared to the Vector group. In shMTF2, the opposite was true. Previous studies have also shown that MTF2 promotes glioma proliferation.32 These results suggest that MTF2 and its interacting genes may influence tumor progression and provide a basis for targeted therapy in future clinical cancer patients.
Conclusions
In summary, a comprehensive pan-cancer analysis of MTF2 reveals the association between MTF2 expression and prognosis, total protein, gene mutations, epigenetic modifications, immune infiltration, and protein-protein interactions in many cancers, leading to a multifaceted under- standing of the role of MTF2 in cancer. This study is based on bioinformatics analysis of a public tumor database with a large sample size, multiple perspectives, high reliability and reference value, and validated by cell biology. However, the study has some limitations. First, most of our studies used only TCGA source data without multiple database validation. Secondly, the results of this study are for phenotypic and functional analysis only. There was no in-depth analysis
of the mechanism. More clarity and underlying data are needed to better assess the potential relationship between MTF2 and tumors.
Author contributions
Xiangmei Cao and Zhongtao Liu conceived the design of the present study. Cui Tang, Ye Lv and Kuihu Ding performed the experiments and data analysis, and contributed to the writing of the manuscript. Yu Cao and Zemei Ma performed the HE and western blot experiments. Qiqi Zhang and Haiyang Zhou performed the database analysis and revised the paper. Lina Yang and Yu Wang revised the paper and checked the data.
Data availability
The data used to support the findings of this study are included within the article.
Acknowledgments
We acknowledge the TCGA and GTEx databases for providing the platform and contribu- tors who uploaded meaningful datasets. We also thank Professor Xiangmei Cao for her help in the process of data analysis, as well as Professor Xiangmei Cao and Professor Zhongtao Liu for the Ningxia Natural Science Foundation to support the experimental funding. National Natu- ral Science Foundation of China (82260509), The National Natural Science Foundation of Ningxia (2022AAC03161 and 2021AAC03375).
Supplementary materials
Supplementary material associated with this article can be found, in the online version, at doi: 10.1016/j.currproblcancer.2023.100957.
References
1. Sung H, Ferlay J, Siegel RL, et al. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J Clin. 2021;71(3):209-249.
2. Siegel RL, Miller KD, Jemal A. Cancer statistics, 2019. CA Cancer J Clin. 2019;69(1):7-34.
3. Blum A, Wang P, Zenklusen JC. SnapShot: TCGA-Analyzed tumors. Cell. 2018;173(2):530.
4. Clough E, Barrett T. The gene expression omnibus database. Methods Mol Biol. 2016;1418:93-110.
5. Tomczak K, Czerwińska P, Wiznerowicz M. The Cancer Genome Atlas (TCGA): an immeasurable source of knowledge. Contemp Oncol (Pozn). 2015;19(1A):A68-A77.
6. Walker E, Manias JL, Chang WY, et al. PCL2 modulates gene regulatory networks controlling self-renewal and com- mitment in embryonic stem cells, Cell cycle (Georgetown. Tex.). 2011;10(1):45-51.
7. Aranda S, Mas G, Di Croce L. Regulation of gene transcription by Polycomb proteins. Sci Adv. 2015;1(11):e1500737.
8. Margueron R, Reinberg D. The polycomb complex PRC2 and its mark in life. Nature. 2011;469(7330):343-349.
9. Walker E, Chang WY, Hunkapiller J, et al. Polycomb-like 2 associates with PRC2 and regulates transcriptional net- works during mouse embryonic stem cell self-renewal and differentiation. Cell Stem Cell. 2010;6(2):153-166.
10. Casanova M, Preissner T, Cerase A, t, et al. Polycomblike 2 facilitates the recruitment of PRC2 Polycomb group complexes to the inactive X chromosome and to target loci in embryonic stem cells. Development. 2011;138(8): 1471-1482.
11. Perino M, van Mierlo G, Karemaker ID, et al. MTF2 recruits polycomb repressive complex 2 by helical-shape-selective DNA binding. Nat Genet .. 2018;50(7):1002-1010.
12. Lamouille S, Xu J, Derynck R. Molecular mechanisms of epithelial-mesenchymal transition. Nat Rev Mol Cell Biol. 2014;15(3):178-196.
13. Dongre A, Weinberg RA. New insights into the mechanisms of epithelial-mesenchymal transition and implications for cancer. Nat Rev Mol Cell Biol .. 2019;20(2):69-84.
14. Wu TT, Cai J, Tian YH, et al. MTF2 induces epithelial-mesenchymal transition and progression of hepatocellular car- cinoma by transcriptionally activating snail. OncoTargets and therapy. 2019; 12:11207-11220.
15. Sun F, Cheng Y, Riordan JD, et al. WDR26 and MTF2 are therapeutic targets in multiple myeloma, J Hematol Oncol, 14 (1), 2021, 203.
16. Maganti HB, Jrade H, Cafariello C, et al. Targeting the MTF2-MDM2 axis sensitizes refractory acute myeloid leukemia to chemotherapy, Cancer Disc, 8 (11), 2018, 1376-1389.
17. Meng X, Zhang Y, Hu Y, et al. LncRNA CCAT1 sponges miR-218-5p to promote EMT, cellular migration and invasion of retinoblastoma by targeting MTF2. Cellular Signalling. 2021;86:110088.
18. Liu J, Lichtenberg T, Hoadley KA, et al. An Integrated TCGA pan-cancer clinical data resource to drive high-quality survival outcome analytics. Cell. 2018; 173(2):400-416.
19. Beroukhim R, Mermel CH, Porter D, et al. The landscape of somatic copy-number alteration across human cancers. Nature. 2010;463(7283):899-905.
20. Thorsson, David L Gibbs, Scott D Brown Vesteinn, et al. The immune landscape of cancer. Immunity. 2018;48(4):812-830.
21. Li T, Fan J, Wang B, et al. TIMER: a web server for comprehensive analysis of tumor-infiltrating immune cells. Cancer Res. 2017;77(21):e108-e110.
22. Zeng D, Ye Z, Shen R, et al, et al. IOBR: multi-omics immuno-oncology biological research to decode tumor microen- vironment and signatures. Front Immunol. 2021;12:687975.
23. Yoshihara K, Shahmoradgoli M, Martínez E, et al. Inferring tumour purity and stromal and immune cell admixture from expression data. Nature Commu. 2013;4:2612.
24. Helm M, Motorin Y. Detecting RNA modifications in the epitranscriptome: predict and validate. Nature Rev Genetics. 2017; 18(5):275-291.
25. Mehdi A, Rabbani SA. Role of methylation in pro- and anti-cancer immunity. Cancers. 2021;13(3):545.
26. Zhang Z, Jones A, Sun CW, et al. PRC2 complexes with JARID2, MTF2, and esPRC2p48 in ES cells to modulate ES cell pluripotency and somatic cell reprogramming, Stem cells (Dayton. Ohio). 2011;29(2).
27. Li H, Liefke R, Jiang J, et al. Polycomb-like proteins link the PRC2 complex to CpG islands. Nature. 2017;549(7671):287-291.
28. Petracovici Ana, Bonasio Roberto. Distinct PRC2 subunits regulate maintenance and establishment of Polycomb re- pression during differentiation. Mol Cell. 2021;81(12):2625-2639.e5.
29. Loh CH, van Genesen S, Perino M, et al. Loss of PRC2 subunits primes lineage choice during exit of pluripotency. Nature Commun. 2021;12(1):6985.
30. Rothberg JLM, Maganti HB, Jrade H, et al. Mtf2-PRC2 control of canonical Wnt signaling is required for definitive erythropoiesis. Cell Disc. 2018;4:21.
31. Maganti HB, Jrade H, Cafariello C, et al. Targeting the MTF2-MDM2 axis sensitizes refractory acute myeloid leukemia to chemotherapy. Cancer Discov. 2018;8(11):1376-1389.
32. Wang F, Gao Y, Lv Y, et al. Polycomb-like 2 regulates PRC2 components to affect proliferation in glioma cells, J Neuro-oncol, 148 (2), 2020, 259-271.
33. Jain P, Ballare C, Blanco E, et al. PHF19 mediated regulation of proliferation and invasiveness in prostate cancer cells. eLife. 2020;9:e51373.
34. Anastasiadi D, Esteve-Codina A, Piferrer F. Consistent inverse correlation between DNA methylation of the first intron and gene expression across tissues and species, Epigenetics Chromatin, 11 (1), 2018, 37.
35. Bewick AJ, Vogel KJ, Moore AJ, et al. Evolution of DNA methylation across insects. Mol Biol Evol. 2017;34(4):654-665.
36. Foret S, Kucharski R, Pellegrini M, et al. DNA methylation dynamics, metabolic fluxes, gene splicing, and alternative phenotypes in honey bees. Proc Natl Acad Sci U S A. 2012;109(13):4968-4973.
37. Fridman WH, Pagès F, Sautès-Fridman C. et al.,The immune contexture in human tumours: impact on clinical out- come. Nature Rev Cancer. 2012;12(4):298-306.
38. Wood SL, Pernemalm M, Crosbie,et PA, et al. The role of the tumor-microenvironment in lung cancer-metastasis and its relationship to potential therapeutic targets. Cancer Treat Rev. 2014;40(4):558-566.
39. Junttila MR, de Sauvage FJ. Influence of tumour micro-environment heterogeneity on therapeutic response. Nature. 2013;501(7467):346-354.
40. Berntsson J, Nodin B, Eberhardet J, et al. Prognostic impact of tumour-infiltrating B cells and plasma cells in colorec- tal cancer. Int J Cancer. 2016; 139(5): 1129-1139.
41. Carvajal-Hausdorf DE, Mani N, Velchetiet V, et al. Objective measurement and clinical significance of IDO1 protein in hormone receptor-positive breast cancer. J Immunother Cancer. 2017;5(1):81.
42. Feng M, Jiang W, Kimet BYS, et al. Phagocytosis checkpoints as new targets for cancer immunotherapy. Nat Rev Cancer. 2019; 19(10):568-586.
43. Rizzo A, Ricci AD, and Brandi G, PD-L1, TMB, MSI, and other predictors of response to immune checkpoint inhibitors in biliary tract cancer, Cancers (Basel), 13 (3), 2021,558.
44. Imperato S, Mistretta C, Marone M, et al. Automodified Poly (ADP-Ribose) polymerase analysis to monitor DNA dam- age in peripheral lymphocytes of floriculturists occupationally exposed to pesticides, Cells, 8 (2), 2019, 137.
45. Otto JE, Kadoch C. A two-faced mSWI/SNF subunit: dual roles for ARID1A in tumor suppression and oncogenicity in the liver. Cancer Cell. 2017;32(5):542-543.
46. Rankin LC, Arpaia N. Treg cells: A LAGging hand holds the double-edged sword of the IL-23 axis. Immunity. 2018;49(2):201-203.
47. hao S, Zhang MH, Zhang Y, et al. The prognostic value of programmed cell death ligand 1 expression in non-Hodgkin lymphoma: a meta-analysis. Cancer Biol Med. 2018; 15(3):290-298.
48. Groudinsky O, Wallis MG, IBousquet, et al. The NAM1/MTF2 nuclear gene product is selectively required for the stability and/or processing of mitochondrial transcripts of the atp6 and of the mosaic, cox1 and cytb genes in sac- charomyces cerevisiae. Mol Gen Genetics: MGG. 1993;240(3):419-427.
49. Nekrasov M, Klymenko T, Fraterman S, et al. Pcl-PRC2 is needed to generate high levels of H3-K27 trimethylation at polycomb target genes. EMBO J. 2007;26:4078-4088.
50. Zhao Y, Ding L, Wang D, et al., EZH2 cooperates with gain-of-function p53 mutants to promote cancer growth and metastasis, EMBO J, 38 (5), 2019, e99599.
51. Chen Q, Cai J, Wang Q, et al. Long noncoding RNA NEAT1, regulated by the EGFR pathway, contributes to glioblastoma progression through the WNT/B-catenin pathway by scaffolding EZH2, Clin Cancer Res: Off J Am Assoc Cancer Res, 24 (3), 2018, 684-695.