Main

Malignant rhabdoid tumor and epithelioid sarcoma are rare tumors of uncertain differentiation as defined by WHO classification.1, 2 Malignant rhabdoid tumor was originally described in 1978 as a rhabdomyosarcomatoid variant of Wilms’ tumor. Malignant rhabdoid tumor arises in varying anatomical locations such as soft tissue, liver, or the central nervous system of infancy or childhood3, 4 while epithelioid sarcoma is a distinctive soft-tissue sarcoma that typically occurs in the distal extremities of adolescents and young adults.1 In 1997, a large cell variant (proximal type) of epithelioid sarcoma was first described.1 Proximal type epithelioid sarcoma often arises in the body trunk, such as the pelvis, perineum, and genital tract of young to middle-aged adults, and often pursues a rather more aggressive clinical course than the so-called conventional type epithelioid sarcoma.1, 4

Epithelioid sarcoma, especially proximal type epithelioid sarcoma, shows rhabdoid features and epithelial differentiation demonstrated by immunohistochemical expression of epithelial markers, such as cytokeratin and epithelial membrane antigen, and loss of SMARCB1 protein expression. These features resemble those of malignant rhabdoid tumor.5, 6 Some data exist regarding the histological and immunohistochemical differences between proximal type epithelioid sarcoma and malignant rhabdoid tumor,7 but such findings are not yet conclusive.

MicroRNAs are small non-coding RNAs of 20–22 nucleotides, typically excised from 60- to 110-nucleotide foldback RNA precursor structures.8 MicroRNAs exert multiple biological functions by negatively regulating the expression of their target genes involved in development, differentiation, apoptosis, and cell proliferation.9 It is suggested that knowledge of microRNA expression profiles in cancer may have substantial value for diagnostic and prognostic determinations, as well as for eventual therapeutic intervention, because the microRNA expression profiles potentially reflect the developmental lineage and differentiation state of cancer.10

In the present study, we analyzed genome-wide microRNA expressions in malignant rhabdoid tumor and epithelioid sarcoma to test whether these sarcomas are distinct tumor entities and that may have a role in the downregulation of SMARCB1 microRNA expression.

Materials and methods

Patients

Malignant rhabdoid tumor and epithelioid sarcoma in the present study were selected from among more than 17 000 cases of bone and soft-tissue tumors registered in the Department of Anatomic Pathology, Graduate School of Medical Sciences, Kyushu University, Fukuoka, Japan between 1978 and 2012. In all, 42 specimens (malignant rhabdoid tumor, 13 cases; proximal type epithelioid sarcoma, 15 cases; conventional type epithelioid sarcoma, 14 cases) were available. These diagnoses were based on light microscopic examination with hematoxylin and eosin staining according to the most recent WHO classification (malignant rhabdoid tumor and epithelioid sarcoma).1, 2 Moreover, immunoperoxidase procedures using the streptavidin-biotin-peroxidase method were carried out in all cases to confirm the diagnosis. In addition, all cases showed a complete loss of SMARCB1 protein expression, and most cases were previously analyzed for SMARCB1 gene alterations by our group.6, 11 Each sample was prepared from a different patient.

MicroRNA microarray analysis was performed in 13 frozen samples (malignant rhabdoid tumor, five cases; proximal type epithelioid sarcoma, two cases; conventional type epithelioid sarcoma, six cases) (Figure 1 and Table 1) and 3 frozen samples of surrounding non-tumorous skeletal muscle that were collected from patients with various types of sarcoma as controls, and quantitative RT–PCR analysis was carried out in all 42 samples: 29 paraffin-embedded materials (8 malignant rhabdoid tumors, 13 proximal type epithelioid sarcomas and 8 conventional-type epithelioid sarcomas) and the 13 frozen samples described above. SMARCB1 gene alteration analyses were not available in three paraffin-embedded malignant rhabdoid tumor cases because of the small amount of material, whereas the remaining nine malignant rhabdoid tumor cases demonstrated SMARCB1 gene alteration causing loss of SMARCB1 protein expression. Two cases of paraffin-embedded proximal type epithelioid sarcoma had SMARCB1 gene alteration causing the loss of SMARCB1 protein expression (Table 2), but the remaining 23 epithelioid sarcoma cases showed no such SMARCB1 gene alterations.

Figure 1
figure 1

Hematoxylin and eosin histologic features and SMARCB1 immunoreactivities in malignant rhabdoid tumor and epithelioid sarcoma cases. (a and b) Malignant rhabdoid tumor case (MRT-5 at Table 1). (c and d) Conventional type epithelioid sarcoma case (CES-4 in Table 1). (e and f) Proximal type epithelioid sarcoma case (PES-3 in Table 1). Immunohistochemically, in the cases with loss of SMARCB1 protein expression, no nuclear expression is observed in tumor cells, whereas infiltrating lymphocytes or vascular endothelial cells show immunoreactivity (b, d and f).

Full size image
Table 1 Profiles of frozen samples using microRNA microarray analysis
Full size table
Table 2 MicroRNA expression values of proximal type epithelioid sarcoma with SMARCB1 gene alteration and these mean values of malignant rhabdoid tumor, proximal type epitheloid sarcoma, and conventional type epithelioid sarcoma
Full size table

MicroRNA Microarray

Total RNAs, including microRNAs, were extracted from frozen and paraffin-embedded samples using the miRNeasy Mini Kit (Qiagen in Japan, Tokyo, Japan) according to the manufacturer’s instructions.

Extracted total RNA was labeled with Hy5 using the miRCURY LNA Array miR labeling kit (Exiqon, Vedbaek, Denmark). Labeled RNAs were hybridized onto 3D-Gene Human miRNA Oligo chips containing 837 anti-sense probes printed in duplicate spots (Toray, Kamakura, Japan). The annotation and oligonucleotide sequences of the probes conformed to the miRBase miRNA database (http://microrna.sanger.ac.uk/sequences/). The chips were carefully washed, and fluorescent signals were scanned with the ScanArray Lite Scanner (PerkinElmer, Waltham, MA) and analyzed using GenePix Pro version 5.0 (Molecular Devices, Sunnyvale, CA). The raw data of each spot were normalized by substitution with the mean intensity of the background signal determined by all blank spots’ signal intensities of 95% confidence intervals. Measurements of both duplicate spots with the signal intensities >2 s.d. of the background signal intensity were considered to be valid. The relative expression level of a given microRNA was calculated by comparing the signal intensities of the averaged valid spots with their mean value throughout the microarray experiments after normalization by their median values adjusted equivalently.

Quantitative RT–PCR

RNAs isolated from frozen or paraffin-embedded materials were reverse transcribed and PCR amplified with the miScript II RT kit and miScript SYBR PCR kit (Qiagen, Valencia, CA, USA) using an ABI Prism 7500 Sequence Detection system (Applied Biosystems) following the manufacturer’s protocols. Quantitations for miR-19a and miR193a-5p were performed using predeveloped miScript reagents (miR-19a, MS00003192; miR193a-5p, MS00008932). A small non-coding RNA gene (SNORD25, MS00014007) was used for normalization. Relative expression levels were calculated following the comparative threshold cycle method: final numerical value=miR-19a or miR193a-5P value/SNORD25 value.

Results

MicroRNA Expression Profiles in Malignant Rhabdoid Tumor and Epithelioid Sarcoma

An unsupervised hierarchical clustering based on the relative expression of microRNAs with valid duplicate spots showed distinctive expression profiles for each sample type, and non-tumorous skeletal muscle, malignant rhabdoid tumor, and epithelioid sarcoma groups examined were properly clustered into the specific tissue type category. However, the microRNA profiles in the proximal type epithelioid sarcoma group and conventional type epithelioid sarcoma group were intermingled (Figure 2).

Figure 2
figure 2

Global microRNA expression and unsupervised hierarchical clustering of malignant rhabdoid tumor (MRT), proximal type epithelioid sarcoma (PES), conventional type epithelioid sarcoma (CES), and non-tumorous skeletal muscle (NSM).

Full size image

To identify the microRNAs that were differentially expressed among malignant rhabdoid tumor, proximal type epithelioid sarcoma and conventional type epithelioid sarcoma, we compared the expression levels of microRNAs among the three groups using ANOVA (P<0.05). As shown in Figure 3, 56 microRNAs were differentially expressed. Three of these 56 microRNAs were identified as microRNAs with the potential of SMARCB1 mRNA inhibition according to the MicroCosm Targets Version 5 (http://www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/v5): namely miR-19a, miR-193a-5p, and miR-486-5p. However, the possibility of a function with INI1 mRNA inhibition was unlikely for miR-486-5p because the mean expression value of miR-486-5p was much higher in non-tumorous skeletal muscle (507.5) than in malignant rhabdoid tumor (32.7), conventional type epithelioid sarcoma (27.3), or proximal type epithelioid sarcoma (59.3) (Table 3).

Figure 3
figure 3

Fifty-six microRNAs showing significant differences between malignant rhabdoid tumor (MRT), proximal type epithelioid sarcoma (PES) and conventional type epithelioid sarcoma (CES) by ANOVA (P<0.05).

Full size image
Table 3 MicroRNA expression ratios of non-tumorous skeletal muscle, malignant rhabdoid tumor, conventional type epitheloid sarcoma, and proximal type epithelioid sarcoma in microarray analysis
Full size table

MiR-19a and miR-193a-5p Expressions

The results of microRNA expressions are summarized in Table 2 and Figure 4. The mean MiR-19a expression values of malignant rhabdoid tumor, proximal type epithelioid sarcoma with no SMARCB1 gene alterations, and conventional type epithelioid sarcoma with no SMARCB1 gene alterations were 10.2, 12.1, and 4.7, respectively (Figure 4a). However, none of the differences were statistically significant (malignant rhabdoid tumor vs proximal type epithelioid sarcoma vs conventional type epithelioid sarcoma, P=0.09; malignant rhabdoid tumor vs proximal type epithelioid sarcoma, P=0.65; malignant rhabdoid tumor vs conventional type epithelioid sarcoma, P=0.12; proximal type epithelioid sarcoma vs conventional type epithelioid sarcoma, P=0.13). The mean MiR-193a-5p expression values of malignant rhabdoid tumor, proximal type epithelioid sarcoma with no SMARCB1 gene alterations, and conventional type epithelioid sarcoma with no SMARCB1 gene alterations were 3.1, 23.7, and 14.6, respectively (Figure 4b). The mean MiR-193a-5p expression values in epithelioid sarcoma with no SMARCB1 gene alterations, both proximal type epithelioid sarcoma and conventional type epithelioid sarcoma, were significantly higher than that in malignant rhabdoid tumor (malignant rhabdoid tumor vs proximal type epithelioid sarcoma vs conventional type epithelioid sarcoma, P=0.002; malignant rhabdoid tumor vs proximal type epithelioid sarcoma, P=0.002; malignant rhabdoid tumor vs conventional type epithelioid sarcoma, P=0.0004; proximal type epithelioid sarcoma vs conventional type epithelioid sarcoma, P=0.16).

Figure 4
figure 4

The expression values of miR-19a (a) and miR193a-5p (b) in malignant rhabdoid tumor, proximal type epithelioid sarcoma with no SMARCB1 gene alterations, and conventional type epithelioid sarcoma with no SMARCB1 gene alterations. There were statistically significant differences among the three groups with respect to expression levels of miR193a-5p (P=0.002) but not miR-19a (P=0.09).

Full size image

The microRNA expression values of the two proximal type epithelioid sarcoma samples with SMARCB1 gene alteration are also exhibited in Table 2. In the two proximal type epithelioid sarcoma samples with SMARCB1 gene alteration, the MiR-193a-5p expression values (4.8, 6.1) were closer to that of malignant rhabdoid tumor (3.1) than to that of the proximal type epithelioid sarcoma samples with no SMARCB1 gene alterations (23.7).

Discussion

The histological features of malignant rhabdoid tumor particularly resemble proximal type epithelioid sarcoma rather than conventional type epithelioid sarcoma; however, several factors distinguishing malignant rhabdoid tumor from epithelioid sarcoma have been reported.4 Clinically, the peak age for the incidence of proximal type epithelioid sarcoma is early adulthood and adolescence; patients younger than 10 years are rare. In contrast, malignant rhabdoid tumor usually occurs in infancy or childhood. Furthermore, the biological behavior of proximal type epithelioid sarcoma is less aggressive than that of malignant rhabdoid tumor. Immunohistochemically, approximately half of all epithelioid sarcoma cases are positive for CD34 and dysadherin, whereas malignant rhabdoid tumor cases reveal no immunoreactivities for these markers.12 By contrast, glypican-3 immunoexpression is found in about half of malignant rhabdoid tumor cases.7 However, microRNA expression profiles in malignant rhabdoid tumor and epithelioid sarcoma have never been investigated.

In terms of microRNA expression profiles of our current study, malignant rhabdoid rumor and epithelioid sarcoma were significantly different, but no significant difference was found between proximal type and conventional type epithelioid sarcoma. Histologically, conventional type epithelioid sarcoma shows a characteristic pseudogranulomatous pattern and a proliferation of eosinophilic epithelioid and spindle-shaped cells exhibiting slight nuclear atypia, vesicular nuclei, and small nucleoli with central necrosis,1 whereas proximal type epithelioid sarcoma shows a multinodular growth pattern and consists of large epithelioid carcinoma-like cells having marked cytological atypia, vesicular nuclei, and prominent nucleoli; these two histologies have obvious differences.1 However, our current results for microRNA expression profiles support the present WHO classification in which conventional type and proximal type epithelioid sarcoma are in the same category, whereas malignant rhabdoid tumor and epithelioid sarcoma are different.1, 2

SMARCB1 is one of the evolutionarily conserved core subunits in the ATP-dependent SWI/SNF complex and is ubiquitously expressed in the nuclei of all normal cells.13, 14 Complete loss of SMARCB1 immunoexpression has been detected in all malignant rhabdoid tumor cases except in rare examples and in almost all epithelioid sarcoma cases.5, 7 About 20% of malignant rhabdoid tumor cases with complete loss of SMARCB1 protein expression have no alterations sufficient to suppress the expression of gene products at either the DNA or RNA level.15, 16 Meanwhile, in proximal type epithelioid sarcoma, there were discrepancies in the frequency of SMARCB1 gene alteration between the investigations of Modena et al (100%), that of Papp et al (19%: rate of biallelic alteration explaining the loss of SMARCB1 function in both types of epithelioid sarcoma), that of Sullivan et al (100%), and that of our previous report (25%).6, 17, 18, 19 Whatever the causes, no mechanism for the inactivation of the SMARCB1 gene product in malignant rhabdoid tumor and epithelioid sarcoma cases with no gene alteration has ever been clarified, and the possibility of regulation by specific microRNA and/or more factors has been considered likely.

The results of our present study suggest that miR-193a-5p can have the role of negatively regulating SMARCB1 mRNA because the mean mi-193a-5p expression value of microarray analysis in epithelioid sarcomas with no SMARCB1 gene alterations (conventional-type; 49.1, proximal type; 44.0) was higher than that in non-tumorous skeletal muscle (13.7), and the miR-193a-5p expression value in epithelioid sarcoma with no SMARCB1 gene alterations was significantly higher than that in malignant rhabdoid tumor (malignant rhabdoid tumor vs proximal type epithelioid sarcoma, P=0.002; malignant rhabdoid tumor vs conventional type epithelioid sarcoma, P=0.0004; proximal type epithelioid sarcoma vs conventional type epithelioid sarcoma, P=0.16). In addition, the miR193a-5p expression values of two cases of proximal type epithelioid sarcoma with SMARCB1 gene alteration (4.8, 6.1) were closer to that of malignant rhabdoid tumor (3.1) than those of epithelioid sarcoma with no SMARCB1 gene alteration (proximal type epithelioid sarcoma, 23.7; conventional type epithelioid sarcoma, 14.6). The function of miR-193a-5p in SMARCB1 mRNA inhibition has not yet been reported. However, it was previously reported that miR-193a-5p might be involved in the mechanism of a predictive tool for ifosfamide response in osteosarcoma or as a target of the YY1-APC regulatory axis in human endometrioid endometrial adenocarcinoma.20, 21 The present study concentrated on analysis in clinical samples, and we did not conduct a verification test using cell lines. Therefore, we cannot confirm absolutely the role of miR-193a-5p.

In summary, we analyzed microRNA expressions in malignant rhabdoid tumor and epithelioid sarcoma. In microarray analysis, conventional type and proximal type epithelioid sarcoma showed almost the same microRNA expression profiles, but both types of epithelioid sarcoma had different profiles from malignant rhabdoid tumor. Accordingly, epithelioid sarcoma, especially proximal type epithelioid sarcoma, and malignant rhabdoid tumor are suggested to be distinct tumors, and both types of epithelioid sarcoma are suggested to be involved in the same category. In addition, the miR193a-5p expression value in malignant rhabdoid tumor was significantly higher than those in proximal type and conventional type epithelioid sarcoma with no SMARCB1 gene alterations (P=0.002). Therefore, it is suggested that miR193a-5p may have the potential to inhibit SMARCB1 mRNA.