The identification of miRNA and mRNA expression profiles associated with pediatric atypical teratoid/rhabdoid tumor

Background Atypical teratoid/rhabdoid tumor (AT/RT) is a malignant pediatric tumor of the central nervous system (CNS) with high recurrence and low survival rates that is often misdiagnosed. MicroRNAs (miRNAs) are involved in the tumorigenesis of numerous pediatric cancers, but their roles in AT/RT remain unclear. Methods In this study, we used miRNA sequencing and gene expression microarrays from patient tissue to study both the miRNAome and transcriptome traits of AT/RT. Results Our findings demonstrate that 5 miRNAs were up-regulated, 16 miRNAs were down-regulated, 179 mRNAs were up-regulated and 402 mRNAs were down-regulated in AT/RT. qPCR revealed that hsa-miR-17-5p and MAP7 mRNA were the most significantly differentially expressed miRNA and mRNA in AT/RT tissues. Furthermore, the results from analyses using the miRTarBase database identified MAP7 mRNA as a target gene of hsa-miR-17-5p. Conclusions Our findings suggest that the dysregulation of hsa-miR-17-5p may be a pivotal event in AT/RT and miRNAs that may represent potential therapeutic targets and diagnostic biomarkers. Supplementary Information The online version contains supplementary material available at 10.1186/s12885-022-09549-6.


Background
Atypical teratoid/rhabdoid tumor is the most common malignant embryonal central nervous system tumor in children below 12 months of age; however, the incidence of AT/RT is not yet known and its incidence rate decreases with age [1,2]. AT/RT was first identified as one of the embryo tumors that represent approximately 1-2% of pediatric intracranial tumors [3]. Because of the lack of clinical manifestation and radiography characteristics, the early clinical diagnosis of AT/RT remains challenging [4,5]. The treatment options for AT/RT currently include surgical resection, chemotherapy and radiotherapy [6][7][8]. However, the prognosis for pediatric patients with AT/RT is still dismal, with a median survival of 15.4 months. Studies have shown that the majority of AT/RT patients show genomic mutations in SMARCB1 (also known as INI1) [9]. However, the precise pathogenesis of this disease is unclear. The identification of novel therapeutics based on the specific mechanism of AT/RT carcinogenesis is therefore critical.
Recent studies have shown that microRNAs (miRNAs) play a vital role in CNS tumorigenesis. MiRNAs, a subtype of small non-coding RNAs, regulate gene expression through recognizing and binding to seed sequence-matching sites in the 3' untranslated regions of target mRNAs [10][11][12]. MiRNAs are involved in the pathogenesis of human malignant tumors and function as oncogenes or Open Access † Xinke Xu, Hongyao Yuan and Junping Pan contributed equally to this work as first authors.
*Correspondence: sjwklfc@126.com tumor suppressors, depending on their downstream targets [13][14][15]. Previous studies have identified abnormal miRNA levels in patients with tumors in the CNS, indicating that miRNAs may play a key role in CNS tumor development [16]. However, knowledge of the miRNA expression profile of AT/RT patients is still limited.
Previous studies revealed that some miRNAs are deregulated in AT/RT, suggesting a role of miRNAs as oncogenes or tumor suppressors [17]. The miRNA expression signatures were found to be associated with bio-molecular and prognostic characteristics in AT/RT. Roles of miRNAs in the regulation of AT/RT, through destabilization or translational repression of target mRNAs, have been widely investigated. However, a comprehensive analysis of the miRNA expression profile in pediatric AT/RT has not yet been conducted.
In this study, we determined the miRNA expression profiles of pediatric AT/RT tumors by analyzing public datasets. Our results revealed that 5 miRNAs were significantly upregulated and 16 miRNAs were significantly downregulated in tumor tissue. Furthermore, we found that abnormal hsa-miR-17-5p expression may play key functions in the tumorigenesis in AT/RT and may represent potential targets for clinical treatment.

Differential expression analysis of GEO datasets
The datasets generated and/or analysed during the current study are available in the GSE42656 (https:// www. ncbi. nlm. nih. gov/ geo/ query/ acc. cgi? acc= GSE42 656) and GSE42657 (https:// www. ncbi. nlm. nih. gov/ geo/ query/ acc. cgi? acc= GSE42 657) repository. were processed by edgeR package in RStudio (version 3.5.0), with a significant cutoff |log2FC|> 2 and P-value < 0.01 [18]. The gene expression profile GSE42656 contains eight control and five AT/RT patients. The miRNA expression dataset is derived from GSE42657, which includes seven control and five tumor tissues. Detailed information of the datasets is listed in Table 1.

Functional analysis
Based on results from the differential expression analysis, we identified the related signaling pathways using Gene Ontology (GO) enrichment analysis. GO terms have three different modules: biological process (BP), molecular function (MF), and cellular component (CC). KEGG pathway analysis was used to identify the significant pathways for dysregulated mRNAs [19][20][21]. GO and KEGG analysis were both used in cluster profiler package in R studio [22]. The P-value was calculated for each enriched function and/ or pathway.

Immunocyte infiltration annotation
We used the CIBERSORT approach to identify inflammatory gene expression signatures in silico to identify the characteristics of the immune response in AT/RT.   CIBERSORT is a computational framework for highthroughput characterization of immune cells [23].

miRNA-mRNA pair analysis
We used miRTarBase to predict the target genes of the differentially expressed miRNAs (http:// mirta rbase. mbc. nctu. edu. tw/) [24]. Differentially expressed genes (DEGs) were extracted and the putative miRNA-mRNA regulatory network was constructed using Cytoscape software (version 3.7.0). To validate the miRNA-mRNA network, we calculated the Pearson values and depicted the correlograms through R software. We evaluated the negative correlation between the key miRNA and target expression.

Reverse transcription quantitative real-time PCR (RT-qPCR)
We used a gene chip to analyze the gene expression profiles. cDNA fragments were purified with a PCR extraction Fig. 1 Differential gene expression analysis of pediatric atypical teratoid tumors. A Heat map depicting gene expression from 13 AT/RT cases and normal brain (columns; ordered automatically by hierarchical clustering). A gradient "heat spectrum" appears at the right; red indicates increased expression, whereas blue denotes decreased levels. B Heat map illustrating the expression of 50 mRNAs. C Heat map illustrating the expression of 21 differentially expressed miRNAs (fold change > 2 and P-value < 0.01) kit (RR037A, Takara, China) following the manufacturer's instructions and then enriched by RT-PCR. Total RNA was extracted using TRIzol reagent (Life Technologies, USA) and quantified using Thermo Nanodrop 2000. RNA (0.5 μg) was subjected to reverse transcription using the Script cDNA Synthesis Kit (RR037A, Takara, China). miRNA and mRNA primer sequences are listed in Tables 2 and 3.

Statistical analysis
Statistical analyses are presented as means ± SD. Comparison between two groups was performed using twotailed Student's t tests, and two-way ANOVAs with general linear model procedures using a univariate approach were applied for more than two groups. All statistical analyses were performed with GraphPad 8 Prism software (San Diego, CA, USA). P < 0.05 was considered statistically significant.

Differential expression profiles for pediatric AT/RT
Using |log fold change|> 2 and P-value < 0.01 as a threshold, we identified a total of 581 DEGs in the tumor group compared with the control group. Among the DEGs, 179 were up-regulated and 402 were down-regulated (Fig. 1A, Supplementary Information.). In addition, 21 differentially expressed miRNAs (DEmiRNAs) were identified, including 5 up-regulated DEmiRNAs and 16 down-regulated DEmiRNAs (Table 4, Fig. 1B-C).

GO enrichment analysis for DEGs
To identify the biological characteristics and signaling pathways involved in the pathogenesis of AT/RT, we next used Clusterprofile in R package to enrich DEGs. The enrichment results of the top 20 genes from the CC, MF and BP categories are shown in Fig. 2. The results indicated that many of the DEGs are closely involved in the formation of synapses. Molecular functions analysis indicated that DEGs were involved in binding to specific molecules, such as growth factor binding, calmodulin binding, and activity of passive membrane transporters. The DEGs were also involved in several critical biological progresses including the regulation of synaptic plasticity, modulation of chemical synaptic transmission, and transportation and secretion of the neurotransmitters, which are all involved in the regulation of nervous system plasticity.

KEGG enrichment analysis for DEGs and the immune infiltration correlation of the expression profile
The KEGG signaling pathway results are shown in Fig. 3. DEGs are highly involved in synaptic function and neurotransmitter transmission. The top enriched pathways include the regulation of "Synaptic vesicle cycle", "GABAergic synapse" and "Glutamatergic synapse, " which are consistent with the results of GO enrichment, indicating that impaired synaptogenesis and synaptic dysfunction could contribute to the formation and clinical manifestation of AT/RT. DEGs were also shown to modulate the "cAMP signaling pathway, " which could affect cell differentiation. We also analyzed the correlation between the expression profile and immune infiltration pathways to identify the association between immune cell types and AT. The proportions of certain immune cells such as memory T cells, resting dendritic cells, neutrophils, and neutrophils were relatively lower in tumor tissues compared with levels in normal tissues. B cells, activated NK cells, and T follicular helper cells showed no difference between tumor tissues and normal tissues.

Construction of the AT/RT-associated miRNA-mRNA correlation and network
To clarify the potential roles of significantly dysregulated miRNAs and to further explore miRNA-mRNA regulatory mechanisms in AT/RT, we identified the potential targets of DEmiRNAs and the genes that were inversely co-expressed with DEmiRNAs using the previously shown gene expression profile. The 581 DEmRNAs and 21 mature DEmiRNAs were analyzed using the miRTarBase database (http:// mirta rbase. mbc. nctu. edu. tw/). A total of 17 DEmiRNAs were found to negatively regulate at least one of the targets in DEm-RNAs. Detailed information for each miRNA-mRNA targeting pair is shown in Table 5. The co-expression network of DEmiRNAs and DEmRNAs was constructed and visualized using Cytoscape software; the results are shown in Fig. 4. miR-17a-5p appeared to play the central role in the DEG network; therefore, miR-17A-5p was selected for further analysis. We next examined the regulatory relationship of miR-17a-5p. The subnetworks shown in Fig. 4 revealed the molecular pathways that were altered by miR-17a-5p. There were 15 mRNAs downregulated by miR-17a-5p. In addition, correlation analysis by Pearson coefficient revealed that KIF5C and DPYSL2 had the highest correlation with miR-17a-5p (Fig. 4, Table 6).

Verification for related mRNA expression levels using qRT-PCR
To investigate the potential function and underlying mechanism of miR-17-5p in AT/RT, we used bioinformatics algorithms and mRNA profiling from AT/RT patients to identify potential target genes of miR-17-5p. The binding of a miRNA to its target mRNA can induce translational silencing or degradation, leading to inhibition or enhancement of gene expression. Various studies have performed expression profiling to identify the roles of miRNAs in AT/RT. Our results indicated that MAP7, PRKCB, CDK1, PPP3R1, CCND1, HDAC1 and CDC20 mRNAs were differentially expressed in AT/RT (Fig. 6). Notably, our study showed that MAP7 plays an important regulatory role in AT/RT.

Discussion
AT/RT is an aggressive pediatric tumor of the CNS. Because of the limited available treatments and poor prognosis of AT/RT, there is an urgent need to identify novel therapeutic targets and develop innovative treatment strategies for this disease [25][26][27]. Mutations and/ or deletions of the SMARCB1 (BAF47/INI1/SNF5) gene are hallmarks of AT/RT tumors, and so far no other recurrent genetic abnormalities have been identified [2,28]. Previous studies showed that HMGA2, LIN28, RPL5, RPL10 and SUN2 are crucial regulators in AT/RT [29,30]. However, the precise molecular mechanism of AT/RT remains largely unknown.
We further found that 179 mRNAs were up-regulated and 402 mRNAs were down-regulated, which could be the result of the dysregulated miRNA networks in AT/ RT, as miRNAs regulate the levels and functions of their target mRNAs. GO analyses revealed that these mRNAs are involved in critical pathways such as the regulation of synaptic plasticity, modulation of chemical synaptic transmission, neurotransmitter transportation and secretion. KEGG pathway analysis showed that "Synaptic vesicle cycle, " "GABAergic synapse" and "Glutamatergic synapse" were related to the DEGs, which is consistent with GO enrichment analysis. These findings suggest that altered synaptogenesis and synaptic dysfunction could contribute to the formation and clinical manifestation of AT/RT. Additionally, DEGs were involved in the canonical pathways such as cAMP signaling pathway, which may contribute to the stemness of the AT/RT tumor cells. Several recent studies have analyzed the influence of the host immune system on cancer prognosis [35]. We performed analyses using CIBERSORT, a computational method for high-throughput characterization of different types of immune cells in complex tissues. Our results demonstrated there was no difference in immune-related cells in AT/RT.