Check for updates

MSI2 expression in adrenocortical carcinoma: Association with unfavorable prognosis and correlation with steroid and immune-related pathways

Luciana C. Veronez1 1 Pablo F. das Chagas2 2 Carolina A. P. Corrêa2 |

Mirella Baroni2 |

Keteryne R. da Silva2 Luis F. Nagano2 Kleiton S. Borges3

Rosane G. P. Queiroz 1 Luiz G. Tone1 Carlos A. Scrideli1

1Department of Pediatrics, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, São Paulo, Brazil

2Department of Genetics, Ribeirão Preto Medical School, University of São Paulo, Ribeirão Preto, São Paulo, Brazil

3Division of Endocrinology, Boston Children’s Hospital, Boston, Massachusetts, USA

Correspondence

Luciana C. Veronez, Department of Pediatrics, Ribeirão Preto Medical School, University of São Paulo, Avenida Bandeirantes, 3900 CEP: 14049-900, Ribeirão Preto/SP, Brazil. Email: lcveronez@usp.br

Funding information

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq); Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP), Grant/Award Numbers: 2014/20341-0, 2017/26160-5, 2018/04477-0; Fundação de Apoio ao Ensino, Pesquisa e Assistência do Hospital das Clínicas da Faculdade de Medicina de Ribeirão Preto da Universidade de São Paulo (FAEPA); Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (Capes), Grant/Award Number: 001

Abstract

Adrenocortical carcinoma (ACC) is a rare, but highly aggressive cancer of the adrenal cortex with a generally poor prognosis. Despite being rare, completely resected ACCs present a high risk of recurrence. Musashi-2 (MSI2) has re- cently been recognized as a potential prognostic biomarker and therapeutic target in many cancers. However, no studies have evaluated the clinical sig- nificance of MSI2 expression in ACC. Here, we addressed MSI2 expression and its association with ACC prognosis and clinicopathological parameters. MSI2 expression was analyzed in TCGA, GSE12368, GSE33371, and GSE49278 ACC datasets; and its correlation with other genes and immune cell infiltration were investigated by using the R2: Genomics Analysis and Visualization Platform and TIMER databases, respectively. Enrichment analysis was per- formed with the DAVID Functional Annotation Tool. Kaplan-Meier curves, log-rank tests, and Cox regression analyses were used to explore the prognostic role of MSI2 in ACC. Our findings demonstrated the potential value of MSI2 overexpression as an independent predictor of poor prognosis in patients with completely resected ACC (hazard ratio 6.715, 95% confidence interval 1.266 - 35.620, p =. 025). In addition, MSI2 overexpression was associated with characteristics of unfavorable prognosis, such as cortisol excess (p =. 002), recurrence (p =. 003), and death (p =. 015); positively correlated with genes related to steroid biosynthesis (p <. 05); and negatively correlated with immune-related pathways (p <. 05). Our findings demonstrate that MSI2 has value as a prognostic marker for completely resected ACC and reinforce the investigation of its role as a possible therapeutic target for patients with ACC.

KEYWORDS

adrenocortical carcinoma, immune infiltration, MSI2 expression, prognosis, steroid biosynthesis

Adrenocortical carcinoma (ACC) is a rare, but highly aggressive cancer of the adrenal cortex with annual in- cidence ranging between 0.7 and 3 cases/million world- wide.1,2 Despite the rarity of ACC, patients present dismal prognoses. The 5-year survival rate is 35%, which decreases to only 15% in patients with advanced tumor stages.1,3 Currently, complete surgical resection is the only curative treatment option for localized disease. For advanced and inoperable ACCs, therapies are limited to the use of the adrenolytic agent mitotane with or without chemotherapy, but effectiveness is poor.2,4 Even after the tumor is completely resected, 50%-70% of the patients tend to develop recurrence or metastasis,3,5 highlighting the urgent need for identifying novel biomarkers that could improve prognosis prediction among this group of patients.

Recently, several studies have shown that dysregu- lated expression and activity of RNA-binding proteins (RBPs) as posttranscriptional regulators are associated with initiation, progression, and chemoresistance of various types of tumors.6-8 The RBP Musashi-2 (MSI2) has recently been recognized as a potential prognostic biomarker or a therapeutic target, or both, in many cancers, including hematopoietic malignancies and solid tumors.6,8 Elevated expression of MSI2 plays important roles in malignant transformation, tumor proliferation, cell cycle, migration, invasion, epithelial-mesenchymal transition (EMT), and autophagy.9-16 However, MSI2 expression and its clinical significance in ACC remain unclear.

Here, we investigated the expression profile of MSI2 and addressed, for the first time, its association with clinical and biological parameters in patients with ACC. Our results showed that higher MSI2 expression was associated with reduced disease-free survival (DFS) and identified as an independent prognostic marker for completely resected ACCs. Also, MSI2 expression was correlated with genes related to steroid biosynthesis and immune-related pathways.

2 MATERIALS AND METHODS |

2.1 Dataset and mRNA expression analysis |

Gene Expression Profiling Interactive Analysis (http://gepia. cancer-pku.cn) was used to compare MSI2 gene expression levels between 79 adult ACCs and 128 non-neoplastic adrenal (NNA) samples; data expression of The Cancer Genome Atlas (TCGA) and the Genotype-Tissue Expression

(GTex) databases were employed.17 Regarding the ACC- TCGA samples, patients’ clinical and demographic data were obtained from the cBioPortal for Cancer Genomics and/or R2: Genomics Analysis and Visualization Platform (http://r2. amc.nl).18,19 Additionally, MSI2 messenger RNA (mRNA) expression was analyzed in independent datasets obtained from the Gene Expression Omnibus (GEO) (GSE12368: 6 NNA, 16 adrenocortical adenoma - ACA, and 12 ACC; GSE33371: 10 NNA, 22 ACA, and 33 ACC; and GSE49278: 44 ACC) using the GEO2R online analysis tool.

2.2 | Identification of functional and pathway enrichment analysis of MSI2-correlated genes

MSI2-correlated genes were identified in the ACC-TCGA dataset by a correlation analysis performed in the R2: Genomics Analysis and Visualization Platform (http://r2. amc.nl) using the false discovery rate as a multiple testing correction, adjusted p-value < . 05 and a correlation coeffi- cient of |0.5| as a cutoff. The biological significance of these genes was explored on the basis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis performed by the DAVID Functional Annotation Tool20 and the Gene Ontol- ogy (GO) enrichment analysis.21 ShinyGO v.06122 was used to summarize the GO terms by removing the redundant terms and constructing a GO Network. The smallest p-value indicates the highest degree of enrichment. Gene expression values of all MSI2-correlated genes were visually represented in a heatmap generated by the Complex- Heatmap R package23 using Pearson’s distance and Ward’s method.

2.3 Correlation analysis of MSI2 expression and immune infiltrates or other genes |

Tumor-infiltrating immune cells in ACC were estimated by using TIMER (http://timer.cistrome.org/), a web tool that estimates the abundances of immune infiltrates by using multiple deconvolution methods based on gene expression data according to xCELL.24 Correlation analysis was per- formed on the basis of Spearman’s correlation and estimated statistical significance. By using ACC-TCGA data, correlation between MSI2 expression, tumor purity, and infiltration le- vels of CD8 and CD4 T cells, B cells, natural killer (NK) cells, macrophages, and dendritic cells were analyzed. Gene ex- pression levels were displayed as log2 TPM. Additionally, correlations between MSI2 expression and genes associated with steroid and aldosterone biosynthesis or gene markers of different immune cells (including CD8 +T cells, T cell

general, B cells, macrophages, NK cells, and different types of functional T cells, such as Th1, Th2, Th17, and Tregs, as well as exhausted T cells) were explored by using Spearman’s correlation in an additional dataset (GSE49278).

2.4 Survival analyses |

Survival was analyzed on the basis of Kaplan-Meier curves and the log-rank test; the median of MSI2 expression in ACC samples was used as a cutoff. Overall survival (OS) was de- fined as the time elapsed between the date of ACC diagnosis and the date of disease-related death (or last follow-up) and was performed in the ACC-TCGA and GSE49278 datasets. To analyze disease-free survival (DFS), only patients who underwent complete surgical resection of the primary tumor (R0) were included; the time elapsed between surgery and the first recurrence, only available in the ACC-TCGA data- set, was considered.

2.5 Statistical analyses |

The association between MSI2 expression and clin- icopathological features of ACCs was analyzed by Mann-Whitney U test in two datasets (ACC-TCGA and GSE49278). Cox proportional hazards regression was used to identify clinical/molecular features that independently im- pact patients’ survival in ACC. For multivariate analysis, only

features with significant values as univariates were included. P-values < . 05 were considered statistically significant. Sta- tistical analyses were performed by using the GraphPad Prism 8 and Statistical Package for Social Sciences (SPSS) software.

3 RESULTS |

3.1 MSI2 expression in ACC |

To investigate whether MSI2 expression is dysregulated in ACC, we analyzed mRNA expression data of publicly available datasets. Compared to NNA controls, ACC samples presented lower MSI2 levels in three in- dependent datasets (ACC-TCGA/GTEx, GSE33371, and GSE12368; Figure 1A-C, respectively). However, in the ACC-TCGA/GTEx data set, this difference did not reach statistical significance (Figure 1A). In addition, in the GSE33371 and GSE12368 cohorts, we observed higher MSI2 expression in adrenocortical adenomas (ACA) compared with ACC (Figure 1B,C).

3.2 | Survival and recurrence predictive value of MSI2 expression

Therefore, we investigated whether MSI2 expression le- vels and patients’ survival were correlated by using the

FIGURE 1 MSI2 expression in ACCs. Comparative analysis of MSI2 expression levels in ACC and non-neoplastic adrenal controls by Mann-Whitney U test in multiple datasets: (A) ACC-TCGA and GTex databases, (B) GSE33171, and (C) GSE12368. Horizontal lines within the boxes correspond to the medians; the upper and lower box limits are the first and third quartiles of the data; and whiskers range from the minimum and maximum values of expression. ACC, adrenocortical carcinoma; GTex, The Genotype-Tissue Expression; NNA, non-neoplastic adrenal; TCGA, The Cancer Genome Atlas

(A)

TCGA/GTEx

(B)

GSE33371

(C)

GSE12368

p=0.006

4.0

3000

MSI2 expression Log2 (TPM + 1)

p<0.001

p<0.001

6

MSI2 normalized expression

p<0.001

MSI2 normalized expression

3.5

2000

4

2

3.0

1000

0

2.5

0

ACC (n=79)

NNA (n=128)

ACC ACA NNA (n=33) (n=22) (n=10)

ACC ACA NNA (n=12) (n=16) (n=6)

FIGURE 2 Impact of MSI2 expression levels on overall and disease-free survival in ACC. Kaplan-Meier plots with the pairwise log-rank test of overall survival using (A) ACC-TCGA (n = 79) and (B) GSE49278 (n =44) datasets; and disease-free survival (C) in the ACC-TCGA dataset according to MSI2 gene expression levels. (D) Disease-free survival estimation by using the combination of MSI2 expression and grade in the ACC-TCGA dataset. Survival analyses were carried out by using the median of gene expression values of ACCs as a cutoff point to determine low- or high-expression groups. DFS analysis included only patients who underwent complete surgical resection of the primary tumor (R0, n = 55). Grade information was only available for 41 R0 tumors. ACC, adrenocortical carcinoma; TCGA, The Cancer Genome Atlas

(A)

ACC-TCGA

(B)

1,0-

1,0

GSE49278

Probability of Overall Survival

0,8-

Probability of Overall Survival

0,8

n=40

n=22

0,6-

0,6-

0,4-

0,4-

n=22

n=39

0,2-

0,2-

MSI2 lower expression

0,0-

MSI2 higher expression

MSI2 lower expression

Log-rank p=0.103

0,0

MSI2 higher expression

Log-rank p=0.262

,00

50,00

100,00

150,00

200,00

,00

50,00

100,00

150,00

200,00

Time (months)

Time (months)

(C)

(D)

1,0-

ACC-TCGA

1,0-

ACC-TCGA

Probability of Disease-free Survival

0,8

Probability of Disease-free survival

0,8

n=19

n=28

0,6-

0,6

n=15

0,4-

0,4-

n=3

n=27

0,2

0,2-

- Low grade/MSI2 lower expression

- High grade/MSI2 lower expression

- Low grade/MSI2 higher expression

MSI2 lower expression

0,0-

MSI2 higher expression

n=4

- High grade/MSI2 higher expression

Log-rank p<0.001

0,0

Log-rank p<0.001

,00

50,00

100,00

150,00

200,00

,00

20,00

40,00

60,00

80,00

100,00

120,00

Time (months)

Time (months)

median of gene expression values in ACC samples as a cutoff point to determine low- or high-expression groups. Kaplan-Meier analyses revealed that MSI2 expression was not significantly associated with OS in the ACC- TCGA or GSE49278 datasets (Figure 2A,B). Next, given that patients have heterogeneous prognosis even after localized ACC is completely resected, we evaluated how MSI2 expression impacted DFS in this subgroup of

patients (degree of resection R0, n = 55) by using the ACC-TCGA cohort. Interestingly, compared with ACC with lower MSI2 levels, higher MSI2 expression was significantly associated with poorer DFS (33.3% vs. 79.4%, p <. 001; Figure 2C). Data regarding the degree of resec- tion were not available in the other datasets.

In the 55 RO patients of the ACC-TCGA cohort, Cox proportional univariate analysis indicated that ACC

recurrence was associated with ENSAT stage, cortisol excess, ACC grade, and MSI2 expression (Table 1). However, multivariate analysis showed that only higher MSI2 expression (hazard ratio [HR] 6.715, 95% con- fidence interval [CI] 1.266-35.620, p =. 025) and high ACC grade (HR 6.129, 95% CI 1.574-23.868, p =. 009) remained independent predictors of reduced DFS (Table 1). Moreover, although the presence of high-grade ACC was associated with a poorer prognosis independent of MSI2 expression levels (Figure 2D), the combination of higher MSI2 expression and low-grade ACCs sig- nificantly indicated worse DFS when compared with the presence of low-grade tumors associated with lower MSI2 expression (Figure 2D). These findings suggested that even when MSI2 expression was analyzed together with other important parameters associated with ACC prognosis, it may be useful to improve the prediction of recurrence among patients with completely resected tumors.

3.3 Correlation of MSI2 expression and clinicopathological features in ACC |

To improve our understanding of the relevance of MSI2 in ACC, we investigated how MSI2 expression levels are associated with clinicopathological characteristics. On the basis of the ACC-TCGA dataset, MSI2 was over- expressed in samples from female patients (p =. 011), functioning tumors (p <. 001), and cortisol-producing

ACCs (p =. 002) as compared with samples from males and non-functional or cortisol-producing tumors, re- spectively (Figure 3 and Table S1). Furthermore, patients who relapsed or died had higher MSI2 expression (p = . 009 and p = . 033) compared to patients who did not suffer from these unfavorable events (Figure 3 and Table S1). In an independent additional cohort (GSE49278), functional ACCs also presented higher MSI2 expression levels (Table S1). These results sug- gested that MSI2 expression patterns might be involved and have potential implications in the clinical status of patients and in the behavior of ACCs.

3.4 Functional and pathway enrichment of MSI2-correlated genes in ACC |

To gain insight into the underlying mechanisms of MSI2 roles in ACC and considering the MSI2 function in gene regulation, we identified genes closely correlated with MSI2 expression by using the ACC-TCGA dataset. This analysis collectively resulted in a total of 662 significantly MSI2- correlated genes (Table S2) using a correlation coefficient of | 0.5| as a cutoff. The expression profiles of all these genes are presented in Figure 4A. Interestingly, pathway and GO en- richment analyses showed that genes positively correlated with MSI2 expression were mainly enriched by processes and pathways involved in steroid and aldosterone bio- synthesis (Figure 4B,C; and Table S3). On the other hand,

TABLE 1 Uni- and multivariate analyses of prognostic factors for disease- free survival in the ACC-TCGA dataset.
VariableUnivariate analysis N HR 95% CI p-valueMultivariate analysis
HR95% CIp-value
Age551.0050.978-1.032.723--
Gender
Female vs. male551.7720.687-4.571.237--
Stage at diagnosis
ENSAT III vs. I-II523.2411.213-8.657.0191.8530.472-7.274.377
Grade
High vs. low557.8972.573-24.240< . 0016.1291.574-23.868.009
Cortisol excess
Present vs. absent512.8541.123-7.256.0281.3720.355-5.304.647
MSI2 expression High vs. low6.0072.009-17.957.0016.7151.266-35.620
55.025

Note: Grade was determined by mitotic counts, where <20 mitoses/50 high-powered fields (HPF) is “Low” and >20/50 HPF is “High.”

Abbreviations: ACC, adrenocortical carcinoma; CI, confidence interval; ENSAT, European Network for the Study of Adrenal Tumors; HR, hazard ratio; N, number of samples; TCGA, The Cancer Genome Atlas.

FIGURE 3 MSI2 expression levels according to clinicopathological features of ACC. MSI2 expression was evaluated comparing female versus male patients, presence versus absence of recurrence, alive versus deceased patients, cortisol-producing versus non-producing ACC, and functioning versus non-functioning tumors (with vs. without hormone production). Horizontal lines within the boxes correspond to the medians; the upper and lower box limits are the first and third quartiles of the data; and whiskers range from the minimum and maximum values of expression. Data analysis was performed by using the ACC-TCGA dataset. ACC, adrenocortical carcinoma; TCGA, The Cancer Genome Atlas

Gender

Recurrence

Outcome

MSI2 normalized expression (Log2)

12

MSI2 normalized expression (Log2)

12

MSI2 normalized expression (Log2)

12

0.011

0.009

0.033

10

10

10

8

8

8

6

6

6

4

4

4

Female (n=48)

Male (n=31)

Absent (n=36)

Present (n=37)

Alive (n=28)

Deceased (n=51)

Cortisol excess

Hormone production

MSI2 normalized expression (Log2)

12

MSI2 normalized expression (Log2)

12

0.002

<0.001

10

10

8

8

&

6

6

4

4

Absent (n=41)

Present (n=33)

Absent (n=26)

Present (n=48)

negatively correlated genes were related to immune- mediated responses (Figures 4B and 4D; and Table 3S). These data reinforced correlations between MSI2 and functional/cortisol-producing ACCs and suggested that MSI2 expression might play a role in the processes of steroid production and immune responses in ACC.

3.5 MSI2 expression and correlation with steroid phenotype and immune infiltrates |

To confirm that MSI2 is associated with a steroid phe- notype in ACCs, we evaluated how MSI2 expression was

correlated with genes that enriched steroid and aldos- terone biosynthesis in our previous KEGG enrichment analysis by using expression data of an additional dataset (GSE49278). MSI2 levels were positively and significantly correlated with CYP21A2, DHCR24, HSD3B2, KCNK3, LDLR, MC2R, NR5A1, PBX1, SCARB1, and SOAT1 ex- pression (Table S4).

By using a statistical method that estimates tumor- infiltrating immune cells from gene expression profile on the basis of the TCGA dataset, we also investigated whether MSI2 expression was correlated with immune cell infiltration levels in ACC. Interestingly, high MSI2 expression levels were positively associated with tumor purity (Figure 5), indicating that MSI2 expression and

FIGURE 4 Functional and pathway enrichment analysis of MSI2-correlated genes in ACC. (A) Heatmap showing the expression profile of the 662 significantly MSI2-correlated genes that were determined by using a correlation coefficient of 10.5l and an adjusted p-value <. 05 as a cutoff. (B) Top 15 KEGG pathways mostly enriched among positive and negatively MSI2-correlated genes in ACC. (C) Gene Ontology terms (biological processes) enrichment among positive and (D) negative MSI2-correlated genes according to ShinyGo. Data analysis was performed by using the ACC-TCGA dataset. ACC, adrenocortical carcinoma; KEGG, Kyoto Encyclopedia of Genes and Genomes; TCGA, The Cancer Genome Atlas

(A)

(B)

MS12

Steroid biosynthesis

Biosynthesis of antibiotics

Aldosterone synthesis and secretion

Terpenoid backbone biosynthesis

Natural killer cell mediated cytotoxicity -

Primary immunodeficiency

Cell adhesion molecules (CAMs)-

Chemokine signaling pathway-

Hematopoietic cell lineage-

z-score

2 1

Antigen processing and presentation -

0

Phagosome

1

Osteoclast differentiation -

2

Viral myocarditis -

T cell receptor signaling pathway-

Staphylococcus aureus infection -

Tuberculosis -

Fc gamma R-mediated phagocytosis

Platelet activation-

Positively correlated genes

NF-kappa B signaling pathway

Negatively correlated genes

T

T

V

T

0

2

4

6

8

10

-log(ajusted p-value)

(C)

(D)

Secondary alcohol metabotic process

Regulation of immune system process

Steroid biosynthetic process Secondary alcohol biosynthetic process

Immune system process

Immune response

Cholesterol metabolic process

Leukocyte activation

Sterol biosynthetic process

Cholesterol biosynthetic process

Lymphocyte activation

Cell activation

Lipid biosynthetic process

T cell activation

Sterol metabolic process

Lipid metabolic process

Positive regulation of immune system process

Regulation of cell activation

Steroid metabolic process

Regulation of immune response

immune cell infiltration in the tumor microenvironment were negatively correlated. Specifically, MSI2 expression showed significant negative correlations with infiltrating levels of CD8 + T cells (Rho =- 0.249, p =. 03), effector memory CD8 + T cells (Rho =- 0.271, p =. 02), dendritic cells (Rho =- 0.376, p =. 001) and activated dendritic cells (Rho =- 0.322, p =. 005), and a positive correlation with B cells (Rho = 0.251, p =. 03) (Figure 5).

To confirm these findings, we investigated how MSI2 expression was correlated with typical gene markers of different types of immune cells (including CD8 + T cells, T cell general, B cells, M1 and M2 macrophages, NK cells, and different types of functional T cells, such as Th1, Th2, Th17, and Tregs, as well as exhausted T cells) by using data of an additional dataset (GSE49278-Table S5). Interestingly, MSI2 expression levels were sig- nificantly moderately correlated with markers of T cells (general-CD2 and ICOS), CD4 T lymphocytes (CD4),

Th17 (IL17A), T cell exhaustion (TIGIT and HAVCR2), NK cells (KIR2DS4, CD48, ITGAL, and KLRK1), and many dendritic cell-related genes (HLA-DQB1, HLA- DRA, NRP1, and ITGAX) (Table S5). Taken together, these results suggested that MSI2 expression correlates with steroid and immune phenotypes in ACC.

4 DISCUSSION |

ACC is a rare, but extremely aggressive endocrine ma- lignancy with a variable risk of recurrence even after the tumor is completely resected. These characteristics de- mand urgent identification of predictive biomarkers of prognosis to guide and improve clinical management.5 Given that aberrant expression of RBPs has been im- plicated in cancer development, and because RBPs have been recognized as biomarkers of prognosis and potential

FIGURE 5 Correlation of MSI2 expression with immune infiltrates in ACC. Estimation of tumor-infiltrating immune cells in the ACC-TCGA dataset by using the web tool TIMER based on gene expression data according to xCELL. Spearman's correlation (Rho) between MSI2 expression, tumor purity, and infiltration levels of subsets of CD8 T cells, dendritic cells, and B cells. Gene expression levels are displayed as log2 TPM. ACC, Adrenocortical carcinoma; TCGA, The Cancer Genome Atlas

MSI2 Expression Level (log2 TPM)

Purity

T cell CD8+_XCELL

T cell CD8+ effector memory_XCELL

6

Rho = 0.433

p = 1.18e-04

Rho = - 0.249

Rho = - 0.271

p = 3.34e-02

p = 2.03e-02

800

4

ACC

2

0

0.2

0.4

0.6

0.8

1.00.00

0.05

0.10

0.15 0.00

0.01

0.02

0.03

0.04

Purity

Infiltration Level

Infiltration Level

MSI2 Expression Level (log2 TPM)

B cell plasma_XCELL

Myeloid dendritic cell_XCELL

Myeloid dendritic cell activated_XCELI

6

Rho = 0.251

p = 3.19e-02

Rho = - 0.376

p = 1.05e-03

Rho = - 0.322

p = 5.48e-03

4

ACC

2

0

1.000 0.005 0.010 0.015 0.020

0.00

0.02

0.04

0.06

0.080.00

0.05

0.10

0.15

0.20

Infiltration Level

Infiltration Level

Infiltration Level

therapeutic targets, here we explored the role of MSI2 expression and its clinical implications in ACC pathobiology.7,8

In this study, we found that MSI2 expression is dys- regulated in multiple cohorts of ACC patients and that MSI2 overexpression is independently correlated with poor DFS in patients with completely resected ACC. This association was confirmed by our findings that MSI2 overexpression is associated with clinical features of unfavorable prognosis such as incidence of death, re- currence, functioning tumors, and cortisol excess. In our analysis, although tumor grade remained the most sig- nificant independent prognostic factor for DFS, around 31% of patients with low-grade ACC still developed re- currence. Within low-grade tumors, we found that lower MSI2 levels were associated with better DFS compared to higher MSI2 levels. This suggests that MSI2 expression may be useful to improve the prediction of recurrence among patients with completely resected ACC. In fact,

recent studies have shown and discussed that the com- bination of clinical biomarkers of prognosis with addi- tional molecular factors appears to improve prognostic stratification when compared to these parameters in- dependently.25,26 However, as data regarding the degree of resection were only available in the ACC-TCGA da- taset, the association of MSI2 and DFS needs to be evaluated in other cohorts.

MSI2 has been shown to regulate important cellular processes in cancer, including cell proliferation, migra- tion, invasion, progression, metastasis, and EMT.27-31 Similar to our findings, increased MSI2 expression has also been correlated with poor prognoses in other tu- mors, such as breast cancer,32 cervical cancer,33 acute lymphoblastic leukemia,34,35 non-small cell lung cancer,36 and pancreatic ductal adenocarcinoma.27 Particularly in cervical cancer, MSI2 expression has also been considered an independent prognostic marker of OS and progression-free survival.33

In our analysis, we also found that MSI2 expression is positively correlated with the expression of genes related to steroid biosynthesis and negatively correlated with chemokine and immune-related pathways. We con- firmed this association when we found higher MSI2 le- vels in cortisol-producing ACC and their negative correlation with immune infiltration levels. Recently, ACC has been recognized as one of the tumors with the highest levels of tumor purity, and consequently with the lowest immune signature and leukocyte fraction, at least in part because of glucocorticoid excess.18,37,38 Emerging evidence has shown that components, activating status, and density of tumor-infiltrating immune cells impact not only patients’ prognosis, but also therapeutic re- sponses, even in the case of ACC.37,39,40 Steroid hor- mones present general immunosuppressive functions and are produced by ovaries, testis, and adrenal cortex.41

Previous studies have shown that MSI2 is required in the development of steroidogenic tissues and is highly expressed in steroidogenic tumors with poor prognoses, such as advanced ovarian cancer.42,43 Moreover, MSI2 and other RBPs, such as HuR, regulate the expression of cytokines and chemokines that mediate the recruitment of many immune cells.44-46 In this context, our results suggest that MSI2 expression may affect both steroid le- vels and immune infiltration, impairing the recruitment of immune cells. Nonetheless, the mechanisms and whether this phenotype is a direct or indirect effect of MSI2 expression or a result of steroid excess have not been explored and require further investigation.

5 CONCLUSION |

For the first time, we have demonstrated that higher MSI2 expression levels are associated with clinical fea- tures of unfavorable prognosis in ACC, and we have shown that MSI2 has prognostic and clinical value in ACC. Furthermore, our findings provide new insight into the possible functions of MSI2 correlated with steroid and immune-related pathways and highlight the im- portant role of RBPs in ACC. Besides implications in prognosis and understanding of ACC biology, identifying markers that might be promising therapeutic targets, as in the case of MSI2, may contribute to refining future treatment strategies.

ACKNOWLEDGMENTS

This study was funded by the São Paulo State Research Foundation (FAPESP - 2014/20341; 2018/04477-0; 2017/ 26160-5), the National Council for Scientific and Tech- nological Development (CNPq), the CAPES Foundation (Financial Code 001), and the Foundation for Support of

Teaching, Research, and Assistance of the University Hospital of the Ribeirão Preto Medical School (FAEPA).

CONFLICT OF INTERESTS

The authors declare that there are no conflicts of interest.

AUTHOR CONTRIBUTIONS

Luciana Chain Veronez, Pablo Ferreira das Chagas, and Carolina Alves Pereira Corrêa reviewed the literature and wrote the manuscript. Luciana Chain Veronez, Pablo Ferreira das Chagas, Carolina Alves Pereira Corrêa, Mirella BaroniM, Keteryne Rodrigues da Silva, and Luis Fernando Nagano performed data analysis and con- ceptualized the figures and tables. Kleiton Silva Borges discussed and critically revised the manuscript. Rosane G. P. Queiroz, Luiz Gonzaga Tone, and Carlos Alberto Scrideli critically revised the text for intellectual content. All the authors read and approved the final version of the manuscript.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

1. Fassnacht M, Dekkers OM, Else T, et al. European Society of Endocrinology Clinical Practice Guidelines on the manage- ment of adrenocortical carcinoma in adults, in collaboration with the European Network for the Study of Adrenal Tumors. Eur J Endocrinol. 2018;179(4):G1-G46. https://doi.org/10. 1530/EJE-18-0608

2. Crona J, Beuschlein F. Adrenocortical carcinoma: towards genomics guided clinical care. Nat Rev Endocrinol. 2019;15(9): 548-560. https://doi.org/10.1038/s41574-019-0221-7

3. Else T, Kim AC, Sabolch A, et al. Adrenocortical carcinoma. Endocr Rev. 2014;35(2):282-326. https://doi.org/10.1210/er. 2013-1029

4. Megerle F, Kroiss M, Hahner S, Fassnacht M. Advanced adrenocortical carcinoma: what to do when first-line therapy fails? Exp Clin Endocrinol Diabetes. 2019;127(2-03):109-116. https://doi.org/10.1055/a-0715-1946

5. Fassnacht M, Terzolo M, Allolio B, et al. Combination che- motherapy in advanced adrenocortical carcinoma. N Engl J Med. 2012;366(23):2189-2197. https://doi.org/10.1056/NEJMoa1200966

6. Kudinov AE, Karanicolas J, Golemis EA, Boumber Y. Musashi RNA-binding proteins as cancer drivers and novel therapeutic targets. Clin Cancer Res. 2017;23(9):2143-2153. https://doi.org/ 10.1158/1078-0432.CCR-16-2728

7. Bish R, Vogel C. RNA binding protein-mediated post- transcriptional gene regulation in medulloblastoma. Mol Cells. 2014;37(5):357-364. https://doi.org/10.14348/molcells.2014.0008

8. das Chagas PF, Baroni M, Brassesco MS, Tone LG. Interplay between the RNA binding-protein Musashi and develop- mental signaling pathways. J Gene Med. 2020;22(1):e3136. https://doi.org/10.1002/jgm.3136

9. Han Y, Ye A, Zhang Y, et al. Musashi-2 silencing exerts potent activity against acute myeloid leukemia and enhances chemo- sensitivity to daunorubicin. PLoS One. 2015;10(8):e0136484. https://doi.org/10.1371/journal.pone.0136484

10. Katz Y, Li F, Lambert NJ, et al. Musashi proteins are post- transcriptional regulators of the epithelial-luminal cell state. eLife. 2014;3:e03915. https://doi.org/10.7554/eLife.03915

11. Sheng W, Dong M, Chen C, Li Y, Liu Q, Dong Q. Musashi2 promotes the development and progression of pancreatic cancer by down-regulating Numb protein. Oncotarget. 2017; 8(9):14359-14373. https://doi.org/10.18632/oncotarget.8736

12. . Sun J, Sheng W, Ma Y, Dong M. Potential role of Musashi-2 RNA-binding protein in cancer EMT. Onco Targets Ther. 2021; 14:1969-1980. https://doi.org/10.2147/OTT.S298438

13. Wang, S, Li N, Yousefi M, et al. Transformation of the in- testinal epithelium by the MSI2 RNA-binding protein. Nat Commun. 2015;6:6517. https://doi.org/10.1038/ncomms7517

14. Wu C, Wyatt AW, Lapuk AV, et al. Integrated genome and transcriptome sequencing identifies a novel form of hybrid and aggressive prostate cancer. J Pathol. 2012;227(1):53-61. https://doi. org/10.1002/path.3987

15. Cox JL, Wilder PJ, Gilmore JM, Wuebben EL, Washburn MP, Rizzino A. The SOX2-interactome in brain cancer cells iden- tifies the requirement of MSI2 and USP9X for the growth of brain tumor cells. PLoS One. 2013;8(5):e62857. https://doi.org/ 10.1371/journal.pone.0062857

16. Yang K, Guo W, Ren T, et al. Knockdown of HMGA2 reg- ulates the level of autophagy via interactions between MSI2 and Beclin1 to inhibit NF1-associated malignant peripheral nerve sheath tumour growth. J Exp Clin Cancer Res. 2019; 38(1):185. https://doi.org/10.1186/s13046-019-1183-2

17. Tang Z, Li C, Kang B, Gao G, Zhang Z. GEPIA: a web server for cancer and normal gene expression profiling and interactive analyses. Nucleic Acids Res. 2017;45(W1):W98-W102. https://doi. org/10.1093/nar/gkx247

18. Zheng S, Cherniack AD, Dewal N, et al. Comprehensive pan- genomic characterization of adrenocortical carcinoma. Cancer Cell. 2016;29(5):723-736. https://doi.org/10.1016/j.ccell.2016.04.002

19. Cerami E, Gao J, Dogrusoz U, et al. The cBio cancer genomics portal: an open platform for exploring multidimensional cancer genomics data. Cancer Discov. 2012;2(5):401-404. https://doi.org/10.1158/2159-8290.CD-12-0095

20. Huang DW, Sherman BT, Lempicki RA. Systematic and integrative analysis of large gene lists using DAVID bioinformatics resources. Nat Protoc. 2009;4(1):44-57. https://doi.org/10.1038/nprot.2008.211

21. Huang DW, Sherman BT, Lempicki RA. Bioinformatics enrichment tools: paths toward the comprehensive func- tional analysis of large gene lists. Nucleic Acids Res. 2009; 37(1):1-13. https://doi.org/10.1093/nar/gkn923

22. Ge SX, Jung D, Yao R. ShinyGO: a graphical gene-set en- richment tool for animals and plants. Bioinformatics. 2020; 36(8):2628-2629. https://doi.org/10.1093/bioinformatics/ btz931

23. Gu Z, Eils R, Schlesner M. Complex heatmaps reveal pat- terns and correlations in multidimensional genomic data. Bioinformatics. 2016;32(18):2847-2849. https://doi.org/10. 1093/bioinformatics/btw313

24. Aran D, Hu Z, Butte AJ. xCell: digitally portraying the tissue cellular heterogeneity landscape. Genome Biol. 2017;18(1):220. https://doi.org/10.1186/s13059-017-1349-1

25. Lacombe A, Soares IC, Mariani B, et al. Sterol O-acyl transferase 1 as a prognostic marker of adrenocortical carcinoma. Cancers. 2020;12(1):27. https://doi.org/10.3390/cancers12010247

26. Lippert J, Appenzeller S, Liang R, et al. Targeted molecular analysis in adrenocortical carcinomas: a strategy toward im- proved personalized prognostication. J Clin Endocrinol Metab. 2018;103(12):4511-4523. https://doi.org/10.1210/jc.2018-01348

27. Guo K, Cui J, Quan M, et al. The novel KLF4/MSI2 signaling pathway regulates growth and metastasis of pancreatic cancer. Clin Cancer Res. 2017;23(3):687-696. https://doi.org/10.1158/ 1078-0432.CCR-16-1064

28. He L, Zhou X, Qu C, et al. Musashi2 predicts poor prognosis and invasion in hepatocellular carcinoma by driving epithelial-mesenchymal transition. J Cell Mol Med. 2014;18(1): 49-58. https://doi.org/10.1111/jcmm.12158

29. Yang C, Zhang W, Wang L, et al. Musashi-2 promotes mi- gration and invasion in bladder cancer via activation of the JAK2/STAT3 pathway. Lab Invest. 2016;96(9):950-958. https:// doi.org/10.1038/labinvest.2016.71

30. 0. Yang Z, Li J, Shi Y, Li L, Guo X. Increased musashi 2 ex- pression indicates a poor prognosis and promotes malignant phenotypes in gastric cancer. Oncol Lett. 2019;17(3):2599-2606. https://doi.org/10.3892/ol.2019.9889

31. Zong Z, Zhou T, Rao L, et al. Musashi2 as a novel predictive biomarker for liver metastasis and poor prognosis in colorectal cancer. Cancer Med. 2016;5(4):623-630. https://doi.org/10. 1002/cam4.624

32. Li M, Li AQ, Zhou SL, Lv H, Wei P, Yang WT. RNA-binding protein MSI2 isoforms expression and regulation in progres- sion of triple-negative breast cancer. J Exp Clin Cancer Res. 2020;39(1):92. https://doi.org/10.1186/s13046-020-01587-x

33. Liu Y, Fan Y, Wang X, Huang Z, Shi K, Zhou B. Musashi-2 is a prognostic marker for the survival of patients with cervical cancer. Oncol Lett. 2018;15(4):5425-5432. https://doi.org/10. 3892/ol.2018.8077

4. Aly RM, Ghazy HF. Prognostic significance of MSI2 predicts unfavorable outcome in adult B-acute lymphoblastic leuke- mia. Int J Lab Hematol. 2015;37(2):272-278. https://doi.org/10. 1111/ijlh.12284

5. Zhao HZ, Jia M, Luo ZB, et al. Prognostic significance of the Musashi-2 (MSI2) gene in childhood acute lymphoblastic leukemia. Neoplasma. 2016;63(1):150-157. https://doi.org/10. 4149/neo_2016_018

36. Kudinov AE, Deneka A, Nikonova AS, et al. Musashi-2 (MSI2) supports TGF-ß signaling and inhibits claudins to promote non-small cell lung cancer (NSCLC) metastasis. Proc Natl Acad Sci U S A. 2016;113(25):6955-6960. https://doi.org/10. 1073/pnas.1513616113

37. Landwehr LS, Altieri B, Schreiner J, et al. Interplay between glucocorticoids and tumor-infiltrating lymphocytes on the prognosis of adrenocortical carcinoma. J Immunother Cancer. 2020;8(1). https://doi.org/10.1136/jitc-2019-000469

38. Thorsson V, Gibbs DL, Brown SD, et al. The immune land- scape of cancer. Immunity. 2018;48(4):812-830. https://doi. org/10.1016/j.immuni.2018.03.023

39. Galon J, Angell HK, Bedognetti D, Marincola FM. The con- tinuum of cancer immunosurveillance: prognostic, predictive, and mechanistic signatures. Immunity. 2013;39(1):11-26. https://doi.org/10.1016/j.immuni.2013.07.008

40. Pagès F, Galon J, Dieu-Nosjean MC, Tartour E, Sautès-Fridman C, Fridman WH. Immune infiltration in human tumors: a prognostic factor that should not be ignored. Oncogene. 2010;29(8):1093-1102. https://doi.org/10.1038/onc.2009.416

41. Bereshchenko O, Bruscoli S, Riccardi C. Glucocorticoids, sex hormones, and immunity. Front Immunol. 2018;9:1332. https://doi.org/10.3389/fimmu.2018.01332

42. Lee J, An S, Choi YM, et al. Musashi-2 is a novel regulator of paclitaxel sensitivity in ovarian cancer cells. Int J Oncol. 2016; 49(5):1945-1952. https://doi.org/10.3892/ijo.2016.3683

43. Sutherland JM, Sobinoff AP, Fraser BA, et al. RNA binding protein Musashi-1 directly targets Msi2 and Erh during early testis germ cell development and interacts with IPO5 upon translocation to the nucleus. FASEB J. 2015;29(7): 2759-2768. https://doi.org/10.1096/fj.14-265868

44. Fan J, Ishmael FT, Fang X, et al. Chemokine transcripts as targets of the RNA-binding protein HuR in human airway epithelium. J Immunol. 2011;186(4):2482-2494. https://doi. org/10.4049/jimmunol.0903634

45. Panganiban RP, Vonakis BM, Ishmael FT, Stellato C. Co- ordinated post-transcriptional regulation of the chemokine

system: messages from CCL2. J Interferon Cytokine Res. 2014; 34(4):255-266. https://doi.org/10.1089/jir.2013.0149

46. Duggimpudi S, Kloetgen A, Maney SK, et al. Transcriptome- wide analysis uncovers the targets of the RNA-binding protein MSI2 and effects of MSI2’s RNA-binding activity on IL-6 sig- naling. J Biol Chem. 2018;293(40):15359-15369. https://doi. org/10.1074/jbc.RA118.002243

SUPPORTING INFORMATION

Additional supporting information may be found in the online version of the article at the publisher’s website.

How to cite this article: Veronez LC, das Chagas PF, Corrêa CAP, et al. MSI2 expression in adrenocortical carcinoma: Association with unfavorable prognosis and correlation with steroid and immune-related pathways. J Cell Biochem. 2021;122:1925-1935.

https://doi.org/10.1002/jcb.30153