Taylor & Francis Taylor & Francis Group
ORIGINAL RESEARCH
Check for updates
Identification of tumor-infiltrating immune cells and prognostic validation of tumor- infiltrating mast cells in adrenocortical carcinoma: results from bioinformatics and real-world data
Xi Tiana,b*, Wenhao Xua,b*, Yuchen Wanga,b*, Aihetaimujiang Anwaiera,b, Hongkai Wanga,b, Fangning Wana,b, Yu Zhua,b, Dalong Caoa,b, Guohai Shia,b, Yiping Zhua,b, Yuanyuan Qua,b, Hailiang Zhanga,b, and Dingwei Yea,b
ªDepartment of Urology, Fudan University Shanghai Cancer Center, Shanghai, China; bDepartment of Oncology, Shanghai Medical College, Fudan University, Shanghai, P.R. China
ABSTRACT
Objective: The purpose of this study was to explore the composition of tumor-infiltrating immune cells (TIIC) and prognostic significance of tumor-infiltrating mast cells (TIMC) in adrenocortical carcinoma (ACC). Methods: The gene expression profiles of ACC were downloaded from the Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GSE90713, GSE12368). The abundance of TIICs in ACC samples was calculated by CIBERSORT algorithm and immunohistochemistry was used to identify mast cells of 39 tumor samples from Fudan University Shanghai Cancer Center (FUSCC). Differentially expressed genes (DEGs) were analyzed by LIMMA package using R software. Survival analysis was analyzed by Kaplan-Meier method and Cox regression models.
Results: The abundance of mast cells (p = . 008) was positively correlated with ACC patients’ outcome in TCGA cohort and was also positively correlated with both overall survival (p < . 05) and progression-free survival (p < . 05) in FUSCC cohort. Different TIMC infiltrations showed significant changes in signaling pathways including DNA replication, nuclear chromosome segregation, and meiotic cell cycle process of ACC. In addition, elevated expression of eight hub genes (KIF18A, CDCA8, SKA1, CEP55, BUB1, CDK1, SGOL1, SGOL2) related to the abundance of TIMC in ACC was significantly correlated with the poor prognosis of the patients.
Conclusion: In conclusion, higher TIMC infiltration was positively correlated with ACC patients’ outcome in both TCGA and FUSCC cohort. Lower TIMC infiltration and elevated expression of hub genes (KIF18A, CDCA8, SKA1, CEP55, BUB1, CDK1, SGOL1, SGOL2) are markedly correlated with aggressive progression and poor prognosis, which might shed lights on novel targets for treatment strategies.
ARTICLE HISTORY
Received 2 April 2020 Revised 11 June 2020 Accepted 12 June 2020
KEYWORDS
Adrenocortical carcinoma; tumor infiltrating immune cells; tumor infiltrating mast cells; CIBERSORT; differentially expressed genes
Introduction
Adrenocortical carcinoma (ACC) is a very rare endocrine malignancy, the Surveillance, Epidemiology, and End Results (SEER) database estimates that the annual incidence rate is about 0.72 per million cases, resulting in 0.2% of all cancer deaths in the United States.1 Due to difficulties in early diagnosis of ACC, most patients have appeared metastasis at first diagnosis, and thus leading to few opportunities for surgery and extremely poor prognosis.2 At present, mitotane remains the only nemesis approved by the U.S. Food and Drug Administration for ACC treatment,3 and high-risk ACC patients can only be treated with mitotane alone or in combination with cytotoxic drugs, while these treatments could merely bring them limited survival benefits.4 In recent years, management of immune microenvironment and immunotherapy development has made a great breakthrough in many cancers, including melanoma, clear cell renal cell carcinoma, non-small cell lung cancer, etc.5,6 Hence, future progress and opportunities in immunotherapy may be able to bring novel treatment managements of ACC.7 Tumor
immune microenvironment plays an important role in the occurrence and progression of cancer and can greatly affect the effect of immunotherapy.8 Therefore, it is of great signifi- cance to evaluate tumor immune microenvironment and investigate major promotion to cancer aggressive progression and unfavorable prognosis of ACC.
In the past few decades, more and more evidences show that the malignant phenotype of tumor is determined by the inherent activity of tumor cells and the complex interaction of various cell types in the tumor microenvironment (TME), especially the tumor-infiltrating immune cells (TIICs).9,10 There is a rising awareness that TIICs can interact with the tumor cells, change the tumor immune microenvironment, and even affect the management of cancer.11,12 Ren et al.13 reported that the interaction between stromal cells and epithelial cells affects the progression of pancreatic cancer and Bingle L et al.14 reported that tumor-related macrophage infiltration has a negative effect on the prognosis of patients with breast and bladder cancer. However, the role TIICs play in ACC remains blurred and it needs to be further studied.
CONTACT Ding-Wei Ye
*Contribute equally
& Supplemental data for this article can be accessed in publisher’s website.
@ 2020 The Author(s). Published with license by Taylor & Francis Group, LLC.
Mast cell is a kind of granulocyte immune cell, which exists in tissues exposed to the external environment. Current studies revealed that the function of tumor-infiltrating mast cell (TIMC) varies greatly with the type of tumor. For instance, Rajput AB et al15 found that matrix-infiltrating mast cells suggest a good prognosis for invasive breast cancer. While Imada A et al.16 claimed that stromal mast cells correlate with angiogenesis and poor outcome in stage I lung adenocarcinoma.
To investigate the differential infiltration amplification of TIICs and elucidate their potential prognostic value in ACC patients, we measured TIICs of ACC by CIBERSORT algo- rithm and constructed interaction networks of differentially expressed genes (DEGs) by LIMMA package using R software. We hypothesized that the possible carcinogenic activity of TIMC may impact multiple hallmarks related to tumor development and lead to poor prognosis of ACC. These findings may reveal potential therapeutic targets and provide insights into the molecular mechanisms of ACC microenvironment.
Materials and methods
An overview of this work
In this work, we firstly used CIBERSRT algorithm to eval- uate the microenvironment of 79 ACC in TCGA cohort. Then, survival analyses were applied to identify the prog- nostic value of tumor-infiltrating immune cells (TIICs), and tumor-infiltrating mast cells (TIMCs) were found positively correlated with patients’ overall survival. In the next step, GSE90713, GSE12368 and FUSCC cohort were used to identify TIMC’s role in microenvironment and prognostic value as external validation. Finally, correlation analysis was applied to evaluate the relationship between TIMCs and other kinds of TIICs. Gene set enrichment analysis and identification of differentially expressed genes were used to explore potential mechanism behind TIMC’s positive prognostic value. A flow chart is drawn in Supplementary figure 1.
Evaluation of TIICs in ACC
The Gene expression profiles and clinical information of patients with ACC were downloaded from the Cancer Genome Atlas (TCGA). Clinical information and transcrip- tome data are combined into a matrix. The CIBERSORT17 analysis tool is a deconvolution algorithm that uses a set of barcode gene expression values (corresponding to a “signature matrix” of 547 genes) to accurately determine the composition of immune cells in tumor sample data. The expression profile was normalized and R software was used to run the CIBERSORT algorithm with the number of permutations was set to 100 to explore the proportion of 22 kinds of TIICs in ACC samples. The bar chart and heat map were drawn to show the composition of TIICs of each sample; correlations between TIICs’ abundance were dis- cussed and a heat map was drawn to display the correlation between cells using pheatmap package.18 TIMC abundance
was equal to the sum of abundance of mast cell resting and mast cell activated.
Survival analysis of TIICs abundance of ACC
After calculating the TIICs’ abundance, we sought to identify whether the kinds of TIICs have prognostic value. Clinical information of patients and TIICs’ abundance data are com- bined into a matrix. According to the median of TIICs’ abun- dance, the samples were divided into high abundance group and low abundance group. R software was applied to draw survival curves to visualize the impact of TIICs’ abundance on patients’ overall survival using survival package.19 TIICs with prognostic value were included in the multivariate Cox regression analysis.
Evaluating microenvironment of ACC and normal tissues from Gene Expression Omnibus (GEO) data sets
The p value of TIMC is the smallest in multivariate Cox regression analysis, so we aim to explore the impact of TIMC on ACC microenvironment in other cohorts. Two chip datasets GSE90713 and GSE12368 were downloaded from GEO (Affymetrix GPL15207 platform, Affymetrix GPL570 platform, respectively). The corresponding genes transformed into a probe were converted into a symbol according to the annotation information on the platform. CIBERSORT was applied to evaluate the microenvironment of ACC and normal tissues.
Validation of TIMCs’ prognostic significance in Fudan university Shanghai cancer center cohort
To further validate TIMCs’ prognostic significance, real- world data were also collected from our institute. This study included 39 ACC patients who underwent surgical treatment from Fudan University Shanghai Cancer Center (FUSCC) between 2013 and 2019, and tumor specimens were obtained with informed consent. Anti-tryptase mono- clonal antibody (Ab2378, diluted 1:10,000; Abcam) was used to identify mast cells using immunohistochemistry (IHC). The positive or negative staining was evaluated by two experienced pathologists and determined as follows. The overall IHC score from 0 to 12 was evaluated accord- ing to the multiplying of the staining intensity and extent score, as previously described.20 The IHC scores 0-3, 4-12 are defined as low TIMC group and high TIMC group. Scatter diagram was drawn to explore the correlations between TIMC abundance with phenotype, and Kaplan- Meier method was applied to validate TIMC’s prognostic significance by comparing two groups survival rates.
Impact of TIMC on ACC microenvironment
Correlation analysis, Gene set enrichment analysis (GSEA), and differential gene expression analysis were applied to explore potential mechanism behind TIMC’s positive prognos- tic value. According to the median value of TIMC abundance, ACC samples were divided into high TIMC group and low
a
C
b
TIMC group. To investigate the impact of TIMC on ACC microenvironment, other types of TIICs’ abundance in two groups were calculated and illustrated. Correlation analysis was applied, and Spearman’rho and p values were calculated to confirm the relationship between TIMC and other kinds of TIICs. GSEA was used to explore the potential involved signal pathways of ACC microenvironment.
Identification of differentially expressed genes (DEGs) related to TIMC
The limma package21 was used to analyze the data (adjusted p-value < 0.05 and fold change of at least 2x) and volcano map was drawn. In the high TIMC group, the genes with increased expression were marked with red, and the genes with decreased expression were marked with blue. Heat map was drawn accord- ing to the expression matrix of the samples to show the difference of gene expression between the two groups. Biological character- istics, such as biological process (BP), molecular functional (MF) and cellular component (CC), were extracted from gene ontology (GO)22 enrichment analysis to determine the role of DEGs in ACC. The pathways associated with DEGs were explored by searching the Kyoto Encyclopedia of Genome and Genome (KEGG)23 which is a database resource for understanding advanced functions and biological systems from large-scale mole- cular data generated by high-throughput experimental techni- ques. Functional enrichment analyses were completed by using ClusterProfiler package.24
Further screening, identification and functional enrichment analysis on DEGs
To find the most important genes which exert subtle impact on TIMC abundance, it is necessary to further screen the DEGs. DEGs were uploaded to the Search Tool for the Retrieval of Interacting Genes (STRING; http://string-db.org) (version 10.0) online database25 to predict protein-protein interaction (PPI) networks of DEGs and analyze the functional interac- tions between proteins. This may help to further understand the potential mechanism of the occurrence and development of ACC. Cytoscape (version 3.5)26 is an open-source
bioinformatics software platform for visualizing molecular interaction networks. MCODE (version 1.4.2),27 a Cytoscape plug-in, was used to find the most significant hub genes with MCODE Score≥5. According to the expression level of the screened hub genes in the high and low TIMC group, the scatter plot was drawn using R, and functional enrichment analyses were carried out.
Survival analysis and functional annotations of hub genes associated with TIMC
Survival analysis and functional annotations of hub genes associated with TIMC were carried out to find potential therapeutic targets. According to the expression level of hub genes, the samples were divided into high expression group and low expression group, and Kaplan-Meier method was applied to evaluate the prognostic value of hub genes. The age, sex, pTNM stage, clinical stage, and expression level of hub genes were brought into the univariate regression analy- sis to find out the factors related to the prognosis of patients. While multivariate Cox regression models, including com- mon prognostic factors (pTNM stage, clinical stage) and expression level of hub genes, were also established to find independent prognostic variables.
Flow cytometry
Flow cytometry (FCM) was used to detect CD8+T and CD4+T cells frequency of CD45+ cells in low and high TIMC samples. Then, some molecules were also detected (IFN-y, IL- 22, IL-17A, IL-4 reveal CD4+T cells’ plasticity and IFN-y, TNF- a, GZMB, PRF1, Ki-67 act as CD8+T cells’ effector molecules). Fresh ACC tissues were obtained from tissue bank of Fudan University Shanghai Cancer Center. Staining of molecules was established by using Fixation/Permeabilization Solution Kit according to the manufacturer’s instructions. The stained cells were washed and re-suspended in the cell staining buffer. FCM was analyzed by BD Celesta and FlowJo software (Tree Star). All the FCM antibodies and reagents are summarized in Supplementary table 2.
Results
Profile of TIICs in ACC microenvironment
According to CIBERSORT algorithm, the proportion of 22 types of TIICs in 79 cases of ACC was calculated and displayed by bar chart. As depicted in Figure 1a-b, the abundance of TIICs in each sample is different, and T cells, natural killer cells, macrophages account for most of the TIICs in ACC microenvironment. Additionally, TIICs’ infiltration levels show strong correlation with each other in ACC microenvironment (Figure 1c). For example, T cells CD8 is positively correlated with macrophages M1 (Correlation coefficient = 0.48) and T cells CD4 memory activated is positively correlated with B cells memory (Correlation coefficient = 0.48)
The prognostic value of TIICs in ACC samples
The samples were divided into high abundance group and low abundance group according to the median of each type of TIICs and survival analysis was carried out. As depicted in Figure 2a-e, the abundance of T cells follicular helper (p = . 003), macrophages M0 (p = .015) were significantly correlated with poor prognosis of ACC. While the abundance of mast cells (p = . 008), B cells naive (p = . 031) and monocytes (p = . 044) were positively associated with patient outcome. Multivariate Cox regression analysis (Table 1) suggested that mast cells were mostly correlated with patients’ outcome with smallest p value (p = . 007). Survival curves of TIICs with no significant prognostic value are shown in Supplementary Figure 2.
T cells follicular helper level
+
high
low
Macrophages MO level
+
high
+
a
b
low
1.00
1.00
Survival probability
0.75
Survival probability
0.75
0.50
0.50
0.25
p=3.011e-03
0.25
p=1.526e-02
0.00
0.00
0
1
2
3
4
5
6
7
8
9
10
11
12
0
1
2
3
4
5
6
7
8
9
10
11
12
Patient at risk
Time(years)
Time(years)
Patient at risk
high
25
24
14
11
8
5
4
3
1
1
0
0
0
high
39
35
21
15
10
8
3
3
2
2
2
1
1
low
54
51
42
31
20
17
9
7
6
4
4
2
2
low
40
40
35
27
18
14
10
7
5
3
2
1
1
0
1
2
3
4
5
6
7
8
9
10
11
12
0
1
2
3
4
5
6
7
8
9
10
11
12
Time(years)
Time(years)
C
Mast cells level
high
+
low
d
B cells naive level 4
high
+ low
1.00-
1.00-
Survival probability
0.75
Survival probability
0.75
0.50
0.50
0.25
p=7.894e-03
0.25
p=3.068e-02
0.00
0.00
0
1
2
3
4
5
6
7
8
9
10
11
12
0
1
2
3
4
5
6
7
8
9
10
11
12
Time(years)
Time(years)
Patient at risk
Patient at risk
high
39
38
31
23
16
14
8
6
3
2
1
1
1
high
39
39
31
25
15
12
8
6
4
3
3
2
2
low
40
37
25
19
12
8
5
4
4
3
3
1
1
low
40
36
25
17
13
10
5
4
3
2
1
0
0
0
1
2
3
4
5
6
7
8
9
10
11
12
0
1
2
3
4
5
6
7
8
9
10
11
12
Time(years)
Time(years)
e
Monocytes level +
high
4
low
1.00
Survival probability
0.75
0.50
0.25
p=4.448e-02
0.00
0
1
2
3
4
5
6
7
8
9
10
11
12
Patient at risk
Time(years)
high
39
37
30
24
17
15
10
8
5
4
3
1
1
low
40
38
26
18
11
7
3
2
2
1
1
1
1
0
1
2
3
4
5
6
7
8
9
10
11
12
Time(years)
| Tumor infiltrating immune cells | HR | pvalue |
|---|---|---|
| B cells naive | 0.003(0-6.240) | 0.134 |
| Monocytes | 0.005(0-6.057) | 0.144 |
| Macrophages M0 | 116.551(1.413-9616.049) | 0.035 |
| Mast cells | 0(0-0.007) | 0.007 |
Microenvironment of ACC and normal tissues from Gene Expression Omnibus (GEO) data sets
As is depicted in Figure 3, the abundance of TIICs varies from sample to sample and TIMC abundance of normal samples is higher than that of tumor tissues in both GSE90713 and GSE12368. Survival analysis cannot be done due to the lack of survival information.
Prognostic significance of TIMC in FUSCC cohort
Clinicopathological characteristics in relation to TIMC infiltra- tion status of 39 ACC patients (Fudan University Shanghai Cancer Center cohort) are listed in Table 2. Low and high magnification views of two groups’ TIMCs are shown in Figure 4a. Higher TIMCs’ abundance was correlated with lower AJCC stage, pT stage, pN stage, pM stage and samples presenting necrosis showed higher TIMCs (Figure 4b-c). In Figure 4d, higher TIMCs’ infiltration was positively correlated with both overall survival (p < . 05) and progression-free survival (p <. 05). In samples which did not present necrosis (Figure 4e), higher TIMCs’ infiltration was still positively correlated with both over- all survival (p < . 05) and progression-free survival (p <. 05).
Microenvironmental changes caused by differential TIMC levels on ACC
Other TIICs’ abundance was different between two groups when the samples were stratified with the median TIMC abundance (Figure 5a). As shown in Figure 5b, there is a positive association
between Macrophage M2 and mast cells (Spearman’rho = 0.334, p = . 003). While B cells naive, Monocytes, T cells follicular helper, Dendritic cells activated and T cells CD4 memory acti- vated are negatively correlated with mast cells (Spearman’rho = - 0.329, p = . 003; Spearman’rho = - 0.321 p = . 004; Spearman’rho = - 0.289, p = . 010; Spearman’rho = - 0.262, p = . 020; Spearman’rho = - 0.239, p = . 034). GSEA was applied to explore the possible cellular explanations of the microenvironmental change and there was a significant difference in gene expression between two groups (Figure 5d). As is depicted in Figure 5e-g, gene expression was significantly enriched in DNA replication, nuclear chromosome segregation, and meiotic cell cycle process in the low TIMC group (Top 20 enriched pathway were listed in Figure 5c). These possible functions are closely associated with activation and proliferation of various immune cells and can cause changes in the abundance of TIICs.
Identification and functional enrichment analysis of DEGs related to TIMC
As shown in the volcanic map, expression of 88 genes decreased and 16 genes increased in the high TIMC group (Figure 6a). The expression level of the 104 genes is differ- ent between two groups (Figure 6b), and these genes are associated with sister chromatid segregation, sister chromatid cohesion, cell cycle, p53, signaling pathway, etc. (Figure 6c-d).
100%
B cells naive
a
B cells memory
Plasma cells
b
80%
T cells CD8
T cells CD4 naive
T cells CD4 memory resting
P<0.05
GSE90713
T cells CD4 memory activated
0.4-
60%
T cells follicular helper
cells regulatory (Tregs)
TIMC abundance
T cells gamma delta
0.3-
NK cells resting
NK cells activated
40%
Monocytes
Macrophages MO
0.2-
Macrophages M1
Macrophages M2
0.1.
20%
Dendritic cells resting
Dendritic cells activated
Mast cells resting
Mast cells activated
0.0
Eosinophils
0%
Neutrophils
Normal tissues Tumor tissues
Normal
Tumor
GSE90713
100%
C
B cells naive
d
B cells memory
Plasma cells
80%
= T cells CD8
GSE12368
T cells CD4 naive
0.20-
T cells CD4 memory resting
P=0.16
T cells CD4 memory activated
60%
T cells follicular helper
TIMC abundance
T cells regulatory (Tregs)
0.15
T cells gamma delta
NK cells resting
NK cells activated
Monocytes
0.10-
40%
Macrophages MO
Macrophages M1
Macrophages M2
0.05
20%
Dendritic cells resting
” Dendritic cells activated
Mast cells resting
Mast cells activated
0.00
Eosinophils
0%
Neutrophils
Normal tissues Tumor tissues
Normal
Tumor
GSE12368
| Characteristics | TIMC infiltration Entire cohort IHC (0-3) IHC (4-12) (N = 39) (N = 22) (N = 17) | x2 | p value | ||
|---|---|---|---|---|---|
| N (%) | |||||
| Age | 0.03 | 0.86 | |||
| <70 years | 34 (87.2) | 19 (86.4) | 15 (88.2) | ||
| ≥70 years | 5 (12.8) | 3 (13.6) | 2 (11.8) | ||
| Gender | 0.22 | 0.64 | |||
| Male | 19 (48.7) | 10 (45.5) | 9 (52.9) | ||
| Female | 20 (51.3) | 12 (54.5) | 8 (47.1) | ||
| AJCC stage | 6.88 | <0.01 | |||
| I-II | 14 (35.9) | 4 (18.2) | 10 (58.8) | ||
| III-IV | 25 (64.1) | 18 (81.8) | 7 (41.2) | ||
| T staget | 2.17 | 0.14 | |||
| T1 - T2 | 20 (51.3) | 9 (40.9) | 11 (64.7) | ||
| T3 - T4 | 19 (48.7) | 13 (59.1) | 6 (35.3) | ||
| N staget | 9.29 | <0.01 | |||
| N0 | 19 (48.7) | 6 (27.3) | 13 (76.5) | ||
| N1 | 20 (51.3) | 16 (72.7) | 4 (23.5) | ||
| M staget | 9.85 | <0.01 | |||
| M0 | 21 (53.8) | 7 (31.8) | 14 (82.4) | ||
| M1 | 18 (46.2) | 15 (68.2) | 3 (17.6) | ||
| Necrosis | 6.88 | <0.01 | |||
| Present | 25 (64.1) | 18 (81.8) | 7 (41.2) | ||
| Absent | 14 (35.9) | 4 (18.2) | 10 (58.8) | ||
t TNM scoring system: Tumor size, Lymph Nodes affected, Metastases. AJCC, American Joint Committee on Cancer.
a
b
C
TIMChigh
127
** , ANOVA P=0.003
12-
TIMCs purity
9
TIMCs purity
9.
6-
6
3
3.
TIMClow
0
Stage I
Stage II
Stage III
Stage IV
0
T1-T2
T3-T4
NO
N1
MO
M1
Necrosisneg
Necrosis008
d
Overall Survival
e
Overall Survival
TIMCslow
TIMCslow
Percent survival (%)
100
n=22
TIMCshigh
Necrosisneg patients
Percent survival (%)
100
n=10
TIMCshigh
:
50-
50-
n=17
n=4
p=0.019, HR(high)=3.258
p=0.018.
(high)=6.569
0
0
All patients
0
20
40
60
80
100
0
20
40
60
80
Time (months)
Time (months)
Progression-free Survival
Progression-free Survival
TIMCslow
TIMCslow
Percent survival (%)
100-
n=22
TIMCshigh
Percent survival (%)
100
TI
n=10
TIMCshigh
1
50
50
n=1
n=4
p=0.034, HR(high)=2.714
p=0.030,
HR(high)=5.420
0
0
0
20
40
60
80
0
20
40
60
80
Time (months)
Time (months)
| Cell Type | Spearman's rho | P value |
|---|---|---|
| Macrophages M2 | 0.334 | 0.003 |
| B cells naive | -0.329 | 0.003 |
| Monocytes | -0.321 | 0.004 |
| T cells follicular helper | -0.289 | 0.010 |
| Dendritic cells activated | -0.262 | 0.020 |
| T cells CD4 memory activated | -0.239 | 0.034 |
| Macrophages MO | -0.222 | 0.050 |
| NK cells resting | -0.181 | 0.110 |
| T cells gamma delta | 0.176 | 0.122 |
| Neutrophils | -0.159 | 0.162 |
| T cells CD4 memory resting | 0.149 | 0.191 |
| Eosinophils | -0.119 | 0.297 |
| NK cells activated | 0.087 | 0.447 |
| Dendritic cells resting | 0.079 | 0.488 |
| T cells CD8 | -0.076 | 0.508 |
| Plasma cells | -0.072 | 0.530 |
| T cells regulatory (Tregs) | -0.071 | 0.537 |
| B cells memory | -0.044 | 0.701 |
| Macrophages M1 | 0.032 | 0.777 |
| T cells CD4 naive | -0.003 | 0.976 |
a
1.0-
Neutrophils
b
Eosinophils
Dendritic cells activated
Dendritic cells resting
Macrophages M2
0.8
Macrophages M1
Macrophages MO
Monocytes
NK cells activated
NK cells resting
0.6
T cells gamma delta
T cells regulatory (Tregs)
T cells follicular helper
T cells CD4 memory activated
T cells CD4 memory resting
0.4-
T cells CD4 naive
T cells CD8
Plasma cells
B cells memory
B cells naive
0.2-
0.0
High TIMC
Low TIMC
C
e
Enrichment plot: GO_DNA_REPLICATION
0.0
Top 20 enriched gene sets in low TIMC infiltration group
Enrichment score (ES)
NES :- 2.046
0.1
GO_MEIOTIC_CELL_CYCLE
NOM-p :< 0.001
GO_MEIOTIC_CHROMOSOME_SEPARATION
1.855
-0.2
FDR-q:0.070
GO_SPINDLE_ORGANIZATION
-1.856
GO_MEIOTIC_CHROMOSOME_SEGREGATION
-1.971
-0.3
-1.972
GO_MITOTIC_SPINDLE_ASSEMBLY
-1972
0.4
GO_ORGANELLE_FISSION
GO_REGULATION_OF_CHROMATIN_BINDING
-1.979
-1.983
0.5
GO_SPINDLE_ASSEMBLY
GO_DNA_DEPENDENT_DNA_REPLICATION
-1.889
0.6
GO_MITOTIC_NUCLEAR_DIVISION
-1.992
GO_MITOTIC_SPINDLE_ORGANIZATION
-1.996
GO_MITOTIC_SISTER_CHROMATID_SEGREGATION
-2.002
-2.004
GO_MEIOTIC_CELL_CYCLE_PROCESS
-2.004
Ranked list metric (Signal2Noise)
GO_CHROMOSOME_SEPARATION
2.017
0.0
GO_SISTER_CHROMATID_SEGREGATION
W” (postively comelated)
GO_DNA_REPLICATION
-2.045
9.4
-2.046
GO_CHROMOSOME_SEGREGATION
.2
GO_ATTACHMENT_OF_SPINDLE_MICROTUBULES_TO_KINETOCH
-2.050
-2.058
Zero eross at 5795
GO_REGULATION_OF_CHROMOSOME_SEGREGATION
-2.068
0.2
GO_NUCLEAR_CHROMOSOME_SEGREGATION
-2.072
0.4
0.0
T (negatively comrelated).
d
q
2.000
4,000
8.000
8.000 10.000 12.000 14,000 10,000 18,000 20,000
high TIMC
Rank in Ordered Dataset
low TIMC
Enrichment profile - Hits
Ranking metric scores
Gene name
f
Enrichment plot: GO_NUCLEAR_CHROMOSOME_SEGREGATION
Enrichment score (ES)
0.0
0.1
NES :- 2.072
-0.2
NOM-p :< 0.001
0.3
FDR-q:0.229
0.4
-0.5
-0.6
Ranked list metric (Signal2Noise)
h’ (positively comeinbed)
O
Zero cross at 5795
14
T (negatively comelased)
·
2,000
4,000
8,000
8,000 10,000 12,000 14,000 16,000 18,000 20,000
Rank in Ordered Dataset
Enrichment profile - Hits
Ranking metric scores
g
Enrichment plot: GO_MEIOTIC_CELL_CYCLE_PROCESS
Enrichment score (ES)
0.0
NES :- 2.004
0.1
NOM-p :< 0.001
-0.2
FDR-q:0.081
0.3
-0.4
0.5
Ranked list metric (Signal2Noise)
0.0
h” (positively correlated)
Zero cross at 5795
0.4
T (negatively comrelated)
a
2,000
4,000
8,000
8,000 10,000 12,000 14,000 18,000 18,000 20,000
Rank in Ordered Dataset
Enrichment profile - Hits
Ranking metric scores
Screening and functional annotations of hub genes
The PPI network among all DEGs was constructed with inter- acted specificity score equal to 0.4 (Figure 7a). The most relevant protein interaction group was selected as hub genes
cluster, including KIF18A, CDCA8, SKA1, CEP55, BUB1, CDK1, SGOL1, SGOL2 (Figure 7b). As shown in Figure 7c, expression of the hub genes was significantly decreased in the high TIMC group compared with low TIMC group (p <. 001).
a
b
Type
High TIMG
Title
Low TIMC
E
6
A
2.0
DOWN
NOT
UP
-$
1.5
-log10(Padj)
1.0
0.5
0.0
-8
-4
0
4
8
Log2FoldChange
C
d
sister chromatid segregation
Cell cycle
sister chromatid cohesion
p53 signaling pathway
forebrain development-
Transcriptional misregulation in cancer-
heterochromatin assembly
Viral carcinogenesis
centromeric sister chromatid cohesion
Oocyte meiosis
chromosome segregation
₹
Cellular senescence
nuclear chromosome segregation
Non-small cell lung cancer-
mitotic sister chromatid segregation
heterochromatin organization
p.adjust
Hepatocellular carcinoma
p.adjust
regulation of cellular senescence
0.005
Melanoma
0.20
0.010
Platinum drug resistance
0.25
spindle
0.015
0.30
0.020
Glioma
0.35
mitotic spindle
Pancreatic cancer
chromosome, centromeric region
Chronic myeloid leukemia
chromosomal region
spindle microtubule
MicroRNAs in cancer
8
midbody
Non-homologous end-joining
Flemming body
Thiamine metabolism
kinetochore
Endocrine resistance
spindle midzone
Progesterone-mediated oocyte maturation
ion channel complex
2-Oxocarboxylic acid metabolism
0.0
2.5
5.0
7.5
10.0
0
1
2
3
4
Functional enrichment analysis showed that the hub genes were mainly enriched in chromosome segregation, chromo- some region, and cell cycle (Figure 7d-e).
Survival analysis of hub genes in ACC
As depicted in Figure 8, the overexpression of hub genes associated with TIMCs abundance is significantly correlated with poor prognosis in ACC patients (p <. 05). Univariate Cox regression model suggested that pTstage, pMstage, clinical stage, and expression of hub genes were significantly correlated with prognosis in (p < . 001; Figure 9a). In the multivariate Cox regression models (Figure 9b-i), the expression of hub genes was still significantly correlated with prognosis in ACC patients (p <. 05).
Mast cell infiltration may exert favorable influence on plasticity of CD8 + T and CD4 + T cells
FCM validated that both CD8 + T and CD4 + T cells’ abun- dance was significantly higher in high TIMC group (Supplementary figure 3A). As is shown in Supplementary figure 3B, IL-4 of CD4 + T cells was significantly higher in high TIMC group. PRF1 and Ki-67 of CD8 + T cells also
elevated significantly in high TIMC group (Supplementary figure 3 C). These findings implied that TIMC might poten- tially promote the accumulation of both CD8 + T and CD4 + T cells and enhanced their plasticity. And this might partly explain that TIMC infiltration act as a favorable prognostic factor.
Discussion
Tumor microenvironment plays an important role in the occurrence and development of cancer. It is composed of various immune cells, cancer-related fibroblasts, and endothe- lial cells,28 and the interaction between these cells also involves kinds of chemokines, cytokin3es, angiogenic mediators, growth factors, and so on.29 Recent evidences suggest that the influence of TIICs is a double-edged sword in cancer,30 and mast cells, T cells follicular helper, macrophages, B cells, monocytes31-34 have been found to have independent prognos- tic significance in different tumors. In this study, we also found that these types of TIICs are significantly associated with the prognosis of ACC patients, which suggests that the tumorigen- esis and progression of ACC are closely related to the TME and shed light on the future research direction.
BAHCCI
a
PRED3
MOH28
MAMY20
b
NING
TMPPE
SLOTAS
00-20
20170
RGBP
KIF18A
ALPEL
VIGP
GPCZ
BERG
GORI
FSCNI
SLC3142
84-80410
NR4AS
ETV4
CAN2
TNFSF13B
OTKI
SUCODAL
CNM404
RRCON
CDAS
SKA1
SGOL1
KAZN
CAMS
CHSTID
SHROOMS TTYMI
CYP2783
HAGS
LMKIB
GSTAL
POMAZ
GSG2
SUCHAS
TFCPZLE
CHOV
CKAPZL
PITX1
OLGAPJ
SOOLE
POUF1
CELSA3
PARZ
CEPSS
VEDCZ
5L2
CDK1
GPRS#
9682
CDCA8
SALLS
SGOL2
AMITEE
MARPA
SLPI
LEF
DOIS
PHPLAS
ZNF3850
SYTLZ
HELLS
LIPZID
DNASEILS
PR337℃
OPPL
I
C50G2
CORS
CEP55
MPPŞ
PNMAZ
FENI
HMGAZ
ALCM11
BUB1
NSMF
GRIMAS
CAONGT
CIONES
CDKNZA
KIAAG224
SPARED
ADAM12
CODG73
POMAGA
SPNS3
NDOU
FAMMICH
PIPSKLI
C
a
b
C
d
61
30
60-
30
MRNA expression
mRNA expression
4
MRNA expression
MRNA expression
“.
20
2
10
0
0
0
0
0
High TIMC
Low TIMC
High TIMC
Low TIMC
High TIMC
Low TIMC
High TIMC
Low TIMC
KIF18A
CDCA8
SKA1
CEP55
e
f
g
h
6
40
10
MRNA expression
4
expression
30
MRNA expression
miRNA expression
4
5
..
820
RNA
2
E10
2
0
0
0
High TIMC
Low TIMC
High TIMC
Low TIMC
High TIMC
Low TIMC
0
High TIMC
Low TIMC
BUB1
CDK1
SGOL1
SGOL2
d
e
chromosome segregation-
mitotic sister chromatid segregation
sister chromatid segregation
Oocyte meiosis-
nuclear chromosome segregation
mitotic nuclear division
mitotic metaphase plate congression
9
metaphase plate congression
Progesterone-mediated oocyte maturation
nuclear division-
establishment of chromosome localization -
chromosome localization
Cell cycle
chromosomal region
chromosome, centromeric region
spindle microtubule
p.adjust
p.adjust
spindle
p53 signaling pathway
mitotic spindle
8
0.005
9.02
kinetochore
0.010
0.04
condensed chromosome outer kinetochore
0.015
midbody
Gap junction
0.06
spindle midzone
0.020
intercellular bridge
microtubule binding
Cellular senescence-
tubulin binding-
histone kinase activity -
microtubule plus-end binding
RNA polymerase II CTD heptapeptide repeat kinase activity-
Viral carcinogenesis
protein serine/threonine kinase activity
5
ATP-dependent microtubule motor activity, plus-end-directed-
cyclin-dependent protein serine/threonine kinase activity-
cyclin binding
Human immunodeficiency virus 1 infection-
cyclin-dependent protein kinase activity-
0
1
2
3
4
5
D
1
2
3
Because of the great variability in the type, grade, or stage of tumor and the distribution of mast cells, it is difficult to clearly define the function of TIMC as pro- or anti-tumorigenic. Current studies revealed that mast cells can mobilize and regulate the activity of T cells, regulatory T cells, and antigen-presenting cells (APC) through soluble media or cell-cell contact.35 In addition to inflammatory effects, mast cells are also thought to synthesize and release effective angiogenic factors, such as VEGF.36 Many studies have shown that TIMC may be related to tumor microves- sels density and promote tumor angiogenesis.37 This study
found that TIMC abundance is a favorable prognostic fac- tor in ACC, while the change of TIMC abundance is accompanied by a wide range of other types of TIICs’ abundance changes which may suggest that TIMC does play a key role in the TME of ACC. In FUSCC cohort, we also found that higher TIMC infiltration was positively correlated with both overall survival and progression-free survival. Thus, TIMC may be a potential therapeutic target. Previous studies have found that cell cycle and chromo- some aberration are very important in ACC tumorigenesis.38 GSEA results show that genes related to
a
Overall Survival
b
Overall Survival
C
Overall Survival
d
Overall Survival
2
Low-BUB1.TPM
High BUB1 TPM
0
LOW KIF18A TPM
High KIF 18A TPM
10
LOW COCA8 TPM
High CDCA8 TPM
9
Low SGOL1 TPM
Logrank p=2.6e-06
Logrank p=3.8e-08
Logrank p=3.90-07
High SGOL1 TPM
Logrank p=4.9c-06
:
HR(high)=7.7
P(HR)=52-05
HR(high)=12
0.8
HR(high)=9.8
P(HR) 2.0-05
Vo
HR(high) 6.8
p(HR) 5.3e-05
Percent survival
nghigh)=38
P(HR) 7.6e-06
Percent survival
n(high)=38
Percent survival
n(high)=38
0.6
n[low)=37
0.6
n(low)=38
0.6
n(low)=38
Percent survival
n|high)=38
:
h@low)=38
0.4
0.4
0.4
04
S
8
S
0.0
00
0.0
:
0
50
100
150
0
50
100
150
0
50
100
150
0
50
100
150
Months
Months
Months
Months
e
Overall Survival
f
Overall Survival
g
Overall Survival
h
Overall Survival
2
LOW CDK1 TPM
0
High CDK1 TPM
Low SGOL2 TPM
High SGOL2 TPM
1
Low CEP55 TPM
9
Low SKA1 TPM
Logrank p=7e-08
Logrank p=0.00011
High CEP55 TPM
Logrank p=5.2e-07
High SKA1 TPM
Logrank p=3.5e-05
0.g
HR(high)=11
HR(high)=4.8
PUHRJ=1.20-05
O.B
p(HR)=0.00042
8
HR(high)=9.6
p(HR)=2e-05
4.8
HR(high)=5.8
n(high)=38
P(HR)=0.00021
Percent survival
n(high)=38
n[low)=38
Percent survival
.n(low)=38
Percent survival
n(high)=38
n(low)=38
Percent survival
n(high)=38
0.6
0.6
0.6
n(low)=38
0.4
0.4
0.4
02
2
0.2
0.2
0.0
0.0
0.0
0
50
100
150
0
50
100
150
0
50
100
150
0
50
100
150
Months
Months
Months
Months
İ
Disease Free Survival
I
Disease Free Survival
k
Disease Free Survival
1
Disease Free Survival
9
Low
9
High BUB1 TPM
Low KIF 18A TPM
0
Low COCA8 TPM
9
High CDCA8 TPM
Low SGOL1 TPM
Logrank p=7.96-06
High KIF18A
High SGOL1 TPM
Logrank p=1.5e-06
Logrank p 3.48-06
Logrank p=0.00028
0.8
HR(high)=5
p(HR)=4.6-05
0.8
HR(high)=5.6
p(HR)=1.4e-05
0.8
HR(high)=5.3
p(HR)=2.5€-05
:
·HR(high)=3.5
Percent survival
Percent survival
Percent survival
Percent survival
P(HR)=0.00058 nghịgh)=38
níhigh)=38
n[low)=37
n(high)=38
Whigh)-38
0.6
0.6
n(low)=38
0.6
n(low)=38
0
n[low]=38
0.4
0.4
0.4
0.4
02
0.2
0.2
8
0.0
8
0.0
8
0
50
100
150
0
50
100
150
D
50
100
150
0
50
100
150
Months
Months
Months
Months
m
Disease Free Survival
n
Disease Free Survival
0
Disease Free Survival
p
Disease Free Survival
2
Low CDK1 TPM
0
Low SGOL2
2
1
High CDK1 TPM
Logrank p=0.00019
High SGOL2 TPM
Low CEP55 TPM
Logrank p=1.2e-05
High CEP55 TPM
Low SKA1 TPM
HR(high)=4.8
Logrank p=1.36-05
High SKA1 TPM
Logrank p=3.86-05
0.8
HR(high)=3.6
p(HR)=0.00043
S
P(HR)=5,4e-05
0
HR(high)=4.6
P(HR)=6e-05
HR(high)=4.3
P(HR)=D.00014
Percent survival
n(high)=38
n(high)=38
nihighy Se
n(low)=38
Percent survival
TƯƠNG=38
Percent survival
n(low)=38
Percent survival
níhigh) 38
0.6
W
0.6
0.6
n(low)=38
0.4
0.4
0.4
à
9
5
3
0
8
0.0
0
50
100
150
0
50
100
150
D
50
100
150
0
50
100
150
Months
Months
Months
Months
cell cycle and chromosome replication are enriched in low TIMC group, indicating that the activation of mast cells is also regulated by them.
In view of the breakthrough progress of immunotherapy in cancer in recent years, whether immunotherapy can be applied to ACC is a problem to be solved.7 The response to immunotherapy usually depends on the interaction between tumor cells and TME, 8,39 so it is of great significance to explore the composition of ACC microenvironment. This study found that there is a significant correlation between TIMC and a variety of TIICs (Macrophage M2,B cells naive, monocytes, T cells follicular helper, Dendritic cells activated and T cells CD4 memory activated) in ACC microenvironment; TIMC can be used as a potential target, which may be beneficial to ACC immunotherapy. We also found that hub genes (KIF18A, CDCA8, SKA1, CEP55, BUB1, CDK1, SGOL1, SGOL2) related to TIMC have inde- pendent prognostic significance. These eight genes may be
used as biomarkers for diagnosis and prognosis and are expected to provide new potential targets for treatment.
Thus, our research has some limitations. The main thing is the retrospective design of our study. Multicenter prospective studies are needed to verify the conclusions. However, because of the rarity of tumor, it is very difficult to carry out rando- mized controlled study in ACC. In addition, there is an urgent need to conduct in vitro and in vivo experiments to reveal the underlying mechanism of mast cell infiltration in ACC.
Conclusion
In conclusion, types of TIICs abundance significantly corre- lated with the prognosis of ACC. Lower TIMC infiltration and elevated expression of hub genes (KIF18A, CDCA8, SKA1, CEP55, BUB1, CDK1, SGOL1, SGOL2) markedly correlated with aggressive progression and poor prognosis, which might shed lights on novel targets for treatment strategies.
| a | pvalue | Hazard ratio | b | pvalue | Hazard ratio | C | pvalue | Hazard ratio | |||
|---|---|---|---|---|---|---|---|---|---|---|---|
| T | 0.047 | 13.162(1.039-166.736) | T | 0.047 | 12.614[1.033-153.992) | ||||||
| age | 0.334 | 1.012(0.988-1.037) | N | 0.111 | 2.871(0.786-10.490) | N | 0.177 | 2.440(0.668-8.917) | |||
| gender | 0.991 | 1.004(0.470-2.145) | M | 0.872 | 0.916(0.315-2.661) | M | 0.962 | 1.026(0,355-2.964) | |||
| stage | 0.458 | 0.396(0.035-4.555) | stage | 0.466 | 0.403(0.035-4.630) | ||||||
| T | <0.001 | 10.980(4.283-28.150) | BUB1 | <0.001 | 2.071(1,390-3.086) | CDCAB | <0.001 | 1.881(1.303-2.717) | |||
| N | 0.107 | 2.221(0.842-5.857) | d | pvalue | Hazard ratio | e | pvalue | Hazard ratio | |||
| T | 0.055 | 10.924(0.948-125.888) | T | 0.017 | 19.198(1.705-216.105) | ||||||
| M | <0.001 | 7.351(3.305-16.351) | N | 0.101 | 2.969(0.808-10.906) | N | 0.018 | 5.158(1.325-20.083) | |||
| M | 0.813 | 0.879(0.303-2.556) | M | 0.855 | 0.907(0.319-2.576) | ||||||
| stage | <0.001 | 7.166(3.027-16.960) | |||||||||
| stage | 0.494 | 0.428(0.037-4.881) | stage | 0.265 | 0.247(0.021-2.880) | ||||||
| SKA1 | <0.001 | 1.544(1.277-1.867) | CDK1 | 0.002 | 1.849(1.244-2.749) | CEP55 | <0.001 | 1.902(1.356-2.659) | |||
| KIF18A | <0.001 | 2.477(1.752-3.504) | f | pvalue | Hazard ratio | g | pvalue | Hazard ratio | |||
| T | 0.032 | 15.494(1 265-189.727) | ₸ | 0.026 | 16.267(1,402-188.791) | ||||||
| SGOL1 | <0.001 | 2.176(1.587-2.982) | N | 0.096 | 3.009(0.822-11.010) | N | 0.150 | 2.581(0.709-9.394) | |||
| M | 0.895 | 1.073(0.375-3.071) | M | 0.782 | 0.859(0.292-2.529) | ||||||
| CDK1 | <0.001 | 2.402(1.691-3.414) | stage | 0.341 | 0.305(0.026-3.526) | stage | 0.412 | 0.360(0.031-4.148) | |||
| KIF18A | 0.001 | 1.851(1 275-2.687) | SGOL1 | 0.002 | 1.773(1.244-2.525) | ||||||
| CDCA8 | <0.001 | 2.287(1.672-3.128) | |||||||||
| h | pvalue | Hazard ratio | i | pvalue | Hazard ratio | ||||||
| SGOL2 | <0.001 | 2.514(1.785-3.542) | ₸ | 0.043 | 13.587[1.084-170.309) | T | 0.009 | 37.607(2.505-564.576) | |||
| N | 0.309 | 1.951(0.538-7.073) | N | 0.399 | 1.754(0.476-6.466) | ||||||
| CEP55 | <0.001 | 2.136(1.616-2.821) | |||||||||
| M | 0.472 | 1.462(0.519-4.122) | M | 0.589 | 1.331(0.472-3.755) | ||||||
| BUB1 | <0.001 | 2.551(1.778-3.660) | stage | 0.418 | 0.365(0.032-4.176) | stage | 0.240 | 0.197(0.013-2.955) | |||
| 0.50 | SGOL2 | <0.001 | 2.100(1.396-3.159) | ------- | SKA1 | <0.001 | 1.777(1.269-2.489) | ||||
| 1.0 150 20 A.D 5.0 | |||||||||||
| Hazard ratio |
Figure 9. The expression of hub genes was significantly correlated with prognosis in both univariate and multivariate cox regression. (a) pTstage, pMstage, clinical stage, and expression of hub genes were significantly correlated with prognosis in univariate regression model (p < . 001). (b-i) In the multivariate cox regression models, the expression of hub genes was still significantly correlated with prognosis (p < . 05).
Abbreviations
TIIC
DEGS
tumor-infiltrating immune cells differentially expressed genes tumor-infiltrating mast cells
TIMC
ACC
adrenocortical carcinoma
TCGA
The Cancer Genome Atlas tumor microenvironment protein-protein interaction
TME
PPI
GSEA
Gene set enrichment analysis
BP
biological process
MF
molecular functional
CC
cellular component
GO
gene ontology
KEGG
FUSCC
Kyoto Encyclopedia of Genome and Genome Fudan University Shanghai Cancer Center
Acknowledgments
We thank the TCGA databases and GEO (ID: GSE90713, GSE12368) for providing ACC gene expression profiles.
Disclosure of potential conflicts of interest
No potential conflicts of interest were disclosed.
Ethics approval and consent to participate
The Ethics approval and consent to participate in the current study were approved and consented by the ethics committee of Fudan University Shanghai Cancer center.
Availability of data and material
The datasets during and/or analyzed during the current study available from the corresponding author on reasonable request.
Authors’ contributions
The work presented here was carried out in collaboration with all authors. YDW, ZHL, and QYY defined the theme of the study and discussed analysis, interpretation, and presentation. TX, XWH, and WYC drafted the manuscript, analyzed the data, developed the algorithm, and explained the results. Aihetaimujiang, WHK and WFN, participated in the collection of relevant data and helped draft the manuscript. ZY and CDL helped to perform the statistical analysis. ZYP and SGH helped revise the manu- script and provided guiding suggestions. All the authors read and approved the final manuscript.
Funding
This work is supported by Grants from the National Key Research and Development Project (No.2019YFC1316000) and National Natural Science Foundation of China (No.81772706 and No.81802525).
References
1. Kebebew E, Reiff E, Duh Q-Y., Clark OH, McMillan A. Extent of Disease at Presentation and Outcome for Adrenocortical Carcinoma: Have We Made Progress ?. 2006. 30(5):872-878.
2. Xu W-H, Wu J, Wang J, Wan F-N, Wang H-K, Cao D-L, Qu -Y-Y, Zhang H-L, Ye D-W. Screening and identification of potential prognostic biomarkers in adrenocortical carcinoma. Front Genet. 2019;10:821. doi:10.3389/fgene.2019.00821.
3. Schteingart DE, Doherty GM, Gauger PG, Giordano TJ, Hammer GD, Korobkin M, Worden FP. Management of patients with adre- nal cancer: recommendations of an international consensus con- ference. Endocr-Relat Cancer. 2005;12(3):667-680.
4. Jasim S, Habra MA. Management of adrenocortical carcinoma. Curr Oncol Rep. 2019;21(3):20. doi:10.1007/s11912-019-0773-7.
5. Xu WH, Xu Y, Wang J, Wan F-N, Wang H-K, Cao D-L, Shi G-H, Qu -Y-Y, Zhang H-L, Ye D-W. Prognostic value and immune infiltration of novel signatures in clear cell renal cell carcinoma microenvironment. Aging (Albany NY). 2019;11(17):6999-7020. doi:10.18632/aging.102233.
6. Hegde PS, Chen DS. Top 10 challenges in cancer immunotherapy. Immunity. 2020;52(1):17-35. doi:10.1016/j.immuni.2019.12.011.
7. Varghese J, Habra MA. Update on adrenocortical carcinoma man- agement and future directions. Curr Opin Endocrinol Diabetes Obes. 2017;24(3):208-214.
8. Petitprez F, Vano YA, Becht E, Giraldo NA, de Reyniès A, Sautès-Fridman C, Fridman WH. Transcriptomic analysis of the tumor microenvironment to guide prognosis and immunotherapies. Cancer Immunol Immunother. 2018;67 (6):981-988. doi:10.1007/s00262-017-2058-z.
9. Candido J, Hagemann T. Cancer-related inflammation. J Clin Immunol. 2013;33 Suppl 1 :S79-S84
10. Swann JB, Smyth MJ. Immune surveillance of tumors. J Clin Immunol. 2007;117(5):1137-1146.
11. Fridman WH, Pagès F, Sautès-Fridman C, Galon J. The immune contexture in human tumours: impact on clinical outcome. Nat Rev Cancer. 2012;12(4):298-306. doi:10.1038/nrc3245.
12. Zitvogel L, Kepp O, Kroemer G. Immune parameters affecting the efficacy of chemotherapeutic regimens. Nat Rev Clin Oncol. 2011;8 (3):151-160. doi:10.1038/nrclinonc.2010.223.
13. Ren B, Cui M, Yang G, Wang H, Feng M, You L, Zhao Y. Tumor microenvironment participates in metastasis of pancreatic cancer. Mol Cancer. 2018;17(1):108. doi:10.1186/s12943-018-0858-1.
14. Bingle L, Brown NJ, Lewis CE. The role of tumour-associated macrophages in tumour progression: implications for new antic- ancer therapies. J Pathol. 2002;196(3):254-265. doi:10.1002/ path.1027.
15. Rajput AB, Turbin DA, Cheang MC, Voduc DK, Leung S, Gelmon KA, Gilks CB, Huntsman DG. Stromal mast cells in invasive breast cancer are a marker of favourable prognosis: a study of 4,444 cases. Breast Cancer Res Treat. 2008;107 (2):249-257. doi:10.1007/s10549-007-9546-3.
16. Imada A. Mast cells correlate with angiogenesis and poor outcome in stage I lung adenocarcinoma. Eur Respir J. 2000;15 (6):1087-1093. doi:10.1034/j.1399-3003.2000.01517.x.
17. Newman AM, Liu CL, Green MR, Gentles AJ, Feng W, Xu Y, Hoang CD, Diehn M, Alizadeh AA. Robust enumeration of cell subsets from tissue expression profiles. Nat Methods. 2015;12(5):453-457.
18. Kolde R. pheatmap: Pretty Heatmaps. 2019.
19. Therneau TM, Grambsch PM, editor. Modeling Survival Data: Extending the Cox Model. In Statistics for Biology and Health. 2000; New York, NY: Springer. p. 39-77.
20. Wang J, Xu W-H, Wei Y, Zhu Y, Qin XJ, Zhang HL, Ye DW. Elevated MRE11 expression associated with progression and poor outcome in prostate cancer. J Cancer. 2019;10(18):4333-4340.
21. Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK. Limma powers differential expression analyses for RNA-sequen- cing and microarray studies. Nucleic Acids Res. 2015;43(7):e47.
22. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat genet. 2000;25(1):25-29.
23. Kanehisa M, Symposium NF, Kanehisa M. The KEGG Database. 2002.
24. Yu G, Wan L-G, Han Y, He Q-Y. clusterProfiler: an R Package for Comparing Biological Themes Among Gene Clusters. OMICS: A Journal of Integrative Biology. 2012;16(5):284-287.
25. Franceschini A, Szklarczyk D, Frankild S, Kuhn M, Simonovic M, Roth A, Lin J, Minguez P, Bork P, Mering VC, et al. STRING v9.1: protein-protein interaction networks, with increased coverage and integration. Nucleic acids research. 2013;41(Database issue):D808- D815.
26. Smoot ME, Ono K, Ruscheinski J, Wang P-L, Ideker T. Cytoscape 2.8: new features for data integration and network visualization. Bioinformatics. 2011;27(3):431-432. doi:10.1093/bioinformatics/ btq675.
27. Bandettini WP, Kellman P, Mancini C, Booker OJ, Vasu S, Leung SW, Wilson JR, Shanbhag SM, Chen MY, Arai AE. MultiContrast Delayed Enhancement (MCODE) improves detection of subendo- cardial myocardial infarction by late gadolinium enhancement cardiovascular magnetic resonance: a clinical validation study. J Cardiovasc Magn Reson. 2012;14:83.
28. Senbabaoglu Y, Gejman RS, Winer AG, Liu M, Van Allen EM, de Velasco G, Miao D, Ostrovnaya I, Drill E, Luna A, et al. Tumor immune microenvironment characterization in clear cell renal cell carcinoma identifies prognostic and immunotherapeutically rele- vant messenger RNA signatures. Genome Biol. 2016;17(1):231. doi:10.1186/s13059-016-1092-z.
29. Bremnes RM, Al-Shibli K, Donnem T, Sirera R, Al-Saad S, Andersen S, Stenvold H, Camps C, Busund L-T. The role of tumor-infiltrating immune cells and chronic inflammation at the tumor site on cancer development, progression, and prognosis: emphasis on non-small cell lung cancer. J Thorac Oncol. 2011;6 (4):824-833. doi:10.1097/JTO.0b013e3182037b76.
30. de Visser KE, Eichten A, Coussens LM. Paradoxical roles of the immune system during cancer development. Nat Rev Cancer. 2006;6(1):24-37. doi:10.1038/nrc1782.
31. Liu Z, Zhu Y, Xu L, Zhang J, Xie H, Fu H, Zhou Q, Chang Y, Dai B, Xu J. Tumor stroma-infiltrating mast cells predict prognosis and adjuvant chemotherapeutic benefits in patients with muscle inva- sive bladder cancer. Oncoimmunology. 2018;7(9):e1474317. doi:10.1080/2162402X.2018.1474317.
32. Gu-Trantien C, Loi S, Garaud S, Equeter C, Libin M, Wind AD, Ravoet M, Buanec HL, Sibille C, Manfouo-Foutsop G, et al. CD4+ follicular helper T cell infiltration predicts breast cancer survival. J Clin Invest. 2013;123(7):2873-2892.
33. Lee H-W, Choi H-J, Ha S-J, Lee K-T, Kwon Y-G. Recruitment of monocytes/macrophages in different tumor microenvironments. Biochim Biophys Acta. 2013;1835(2):170-179.
34. Tadmor T, Zhang Y, Cho H-M, Podack ER, Rosenblatt JD. The absence of B lymphocytes reduces the number and function of T-regulatory cells and enhances the anti-tumor response in a murine tumor model. Cancer Immunol Immunother. 2011;60 (5):609-619. doi:10.1007/s00262-011-0972-z.
35. Rigoni A, Colombo MP, Pucillo C. The Role of Mast Cells in Molding the Tumor Microenvironment. Cancer Microenviron. 2015;8(3):167-176.
36. Crivellato E, Nico B, Ribatti D. Mast cells and tumour angiogen- esis: New insight from experimental carcinogenesis. Cancer letters. 2008;269(1):1-6.
37. Sari A, Calli A, Cakalagaoglu F, Altınboga AA, Bal K. Association of mast cells with microvessel density in urothelial carcinomas of the urinary bladder. Ann Diagn Pathol. 2012;16(1):1-6.
38. Giordano TJ, Kuick R, Else T, Gauger PG, Vinco M, Bauersfeld J, Sanders D, Thomas DG, Doherty G, Hammer G. Molecular classi- fication and prognostication of adrenocortical tumors by transcrip- tome profiling. Clin Cancer Res. 2009;15(2):668-676.
39. Wood SL, Pernemalm M, Crosbie PA, Whetton AD. 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.