Low tissue levels of miR-125b predict malignancy in solitary fibrous tumors of the pleura

Solitary fibrous tumors of the pleura (SFTP) are rare neoplasia within the chest cavity with an estimated incidence of approximately 0.2/100’000 persons per year [1], and account for 5% of all pleural tumors [2]. Beside the preferred occurrence in the pleura diverse anatomical locations of solitary fibrous tumors have been reported [3, 4]. Historically, SFTP were denominated as “localized mesothelioma” reflecting the lack of consensus with respect to pathogenesis and biological behavior of these tumors [5]. Nowadays, SFTP have been shown to arise from mesenchymal cells and, as such, express immuno-histochemical markers including vimentin and CD34 [6, 7]. The majority of SFTP follow a benign course with a high cure rate after complete surgical resection [8, 9]. However, malignant variants with a high risk of local recurrence, metastases, and increased mortality are found in 7–60% [9‐11]. In addition, some SFTP which were initially assessed to follow a benign course may transform into malignant variants, sometimes several years after R0-resection [12, 13]. Therefore, characteristics that distinguish benign from malignant variants are crucial for estimating the risk of an adverse outcome, and planning adjuvant therapies and adequate follow-up examinations [5].

England et al. were the first to describe six clinico-pathologic features predicting a malignant behavior. These risk factors included tumor size, localization, sessile tumor, existence of necrosis or hemorrhage, high mitoses count, and evidence of nuclear pleomorphism [9]. More recently, a staging system based on tumor seating and on the above mentioned features have been proposed in order to improve identification of malignant variants [14]. However, several multivariable models identified these characteristics which are based exclusively on clinico-pathologic features as poor predictors for the biological behavior of SFTP [15‐17]. Yet, there is insufficient data on immune-histochemical and molecular markers predicting the outcome of SFTP.

MiRNAs are small non-coding RNA fragments of 20–22 nucleotides in length that are commonly associated with gene silencing. Since there is growing evidence that miRNAs might be of major interest in the diagnosis and prognosis of different diseases including lung cancer [18] and mesothelioma [19], this study aimed to investigate the role of miRNAs as biomarkers to predict the clinical behavior of SFTP.

In total, 38 patients (40% male, mean age 62.2 (±10.9) years) were included (Fig. 1). Baseline characteristics as well as the traditional pathological features of SFTP published by England et al. [9] are summarized in Table 3. Sixteen patients were found to have malignant variants according to completion of two, three, or four to five of these features in seven, four and five patients, respectively.

In this translational study, we addressed the pattern and expression levels of miRNAs in SFTP and found that i) reduced levels of miR-125b are indicative of malignant transformation of SFTP and, ii), lower expression of miR-125b was inversely correlated with tumor size.

The challenge of diagnosis and staging of SFTP is the prediction of biological behavior. The majority of SFTP follow a benign course, but a considerable high proportion of these tumors may present with an aggressive behavior including distant metastases or recurrence, sometimes several years after complete resection [9‐13, 22, 23]. For this reason, several scoring systems have been introduced which are originally based on six pathological features proposed by England et al. [9, 14]. However, multivariable models identified these features as poor predictors for the biological behavior of SFTP [15‐17]. And, clinical parameters do not seem to have sufficient discriminative power to predict the outcome of SFTP [17]. Thus, investigation on immuno-histochemical and molecular markers is needed to characterize SFTP and their outcome more accurately. For example, high p53 expression has been shown to be associated with adverse outcome of SFTP [15, 24]. Recently, we found that Ki-67 labeling index (MIB-1), an immuno-histochemical marker of proliferation, is a promising predictor for an adverse outcome in SFTP [17]. Due to its role in ribosomal RNA transcription [25], the Ki-67 antigen is of special interest in the context of non-coding RNAs profiling. These small RNA fragments have emerged as important gene regulators and it is estimated that more than 50% of the human genome are controlled by non-coding RNAs [26]. Within the group of non-coding RNAs, miRNAs are the best-characterized to date [27]. As such, miRNAs have been implicated in the pathogenesis of disease and as biomarkers for diagnosis and prognosis. Of interest, miRNAs have distinct properties that make them interesting biomarker candidates including stability and standardized quantification within tissues and biofluids [28].

Here we employed a screening approach and assessed the expression levels of selected miRNAs in clinically well-characterized samples of SFTP. Isolation and quantification of miRNAs was performed in formalin-fixed and paraffin-embedded (FFPE) tissue blocks. This method has been described before [29, 30] and was recently confirmed as a feasible technology by our own work using lung tissue slides [31]. These data, together with the results provided here, highlight both the reliability and the reproducibility by which miRNAs are isolated and quantified from FFPE.

Of interest, out of a panel of selected miRNAs that have previously been associated with lung cancer [18], only the expression of miR-125b was found to be significantly altered when benign and malignant variants of SFTP were compared. miR-125b has been identified as an important oncomir [32] that is dysregulated in many different types of cancers including breast cancer [33], hepatocellular and cologastric carcinoma [34] and thyroid neoplasias [35]. Since both up and down regulation of miR-125b have been described in cancer, it was suggested that miR-125b has tissue or cell-type specific effects [32]. In this context, miR-125b was linked to aberrant regulation of several target genes involved in the control of differentiation, proliferation and apoptosis [32]. Whether and which of these mechanisms apply to SFTP is beyond the scope of this descriptive analysis and remains unclear to date. However, since tumor size was significantly correlated with expression levels of miR-125b, our results here provide evidence that, in addition to mediating malignant transformation of SFTP, miR-125b might have pro-proliferative and anti-apoptotic effects on pleural cells.

Our study is mainly limited by its descriptive design and by a relatively low patient number. However, by analyzing a substantial number of samples obtained from two different tertiary care centers, this is the first report on a pathogenetic role of miRNAs in SFTP. Furthermore, since there are concerns regarding the accuracy of the traditional grading systems of SFTP [15‐17], miR-125b should be prospectively investigated in a future study with higher sample size. Moreover, the discriminative power of miR-125b should be confirmed with outcome data and other immunohistochemical parameters (e.g. p53 expression or MIB-1 rather than traditional grading systems. However, due to the low sample size we were not able to overcome this limitation.

Our data show that lower miR-125b expression levels are associated with an increase in tumor size and that downregulation of miR-125b indicates malignant transformation. These data emphasize a role of miR-125b in the clinical course of SFTP. Further studies have to address the potential use of miR-125b as a biomarker or therapeutic agent in SFTP.