Circulating miR-21-5p and miR-148a-3p as emerging non-invasive biomarkers in thymic epithelial tumors

ABSTRACT

Thymic epithelial cells give rise to both thymoma and thymic carcinoma. A crucial advance in thymic epithelial tumors (TET) management may derive from the identification of novel molecular biomarkers able to improve diagnosis, prognosis and treatment planning.In a previous study, we identified microRNAs that were differentially expressed in tumor vs normal thymic tissues. Among the microRNAs resulted up-regulated in TET tissues, we evaluated miR-21-5p, miR-148a-3p, miR-141-3p, miR-34b-5p, miR-34c-5p, miR-455-5p as blood plasma circulating non-invasive biomarkers for TET management.We firstly report that the expression levels of specific onco-miRNAs, that we found upregulated in the blood plasma collected from TET patients at surgery, resulted significantly reduced in follow-up samples.This pilot study suggests that circulating miR-21-5p and miR-148a-3p could represent novel non-invasive biomarkers to evaluate the efficacy of therapy and the prognosis of TET.

Introduction

Thymic epithelial cells give rise to both thymoma and thymic carcinoma. These thymic epithelial tumors (TET), although considered rare neoplasms, represent the most frequent tumors in the anterior mediastinal compartment in adults.¹ TET can occur at all ages, with a peak of frequency around 30-40 or 60-70 years of age in thymomas associated or not, respectively, with Myasthenia Gravis (MG).^1,2 The evolution of the disease is often unpredictable, ranging from an indolent attitude to multiple local relapses and, more rarely, to extrathoracic spread.² The modern histological classifications of TET identify 5 major thymoma subtypes. Although Type A, Type AB and Type B1 thymomas have an excellent prognosis when discovered in initial stages, Type B2 and especially Type B3 thymomas, together with thymic carcinomas, show an unfavorable outcome, particularly when discovered at advanced stages.^3,4

A crucial advance in TET management could derive from the identification of new molecular biomarkers related to these neoplasms able to improve diagnosis, prognosis and treatment planning.^5,6

MicroRNA expression is tissue-specific and highly regulated according to the cells developmental lineage and stage.⁷ Several lines of evidence indicate that miRNAs might be differentially expressed in cancer cells, in which they form unique expression patterns or signatures.⁸ Specific miRNA-expression signatures are identified in correlation with tumor classification and are also emerging as predictive markers of lymph node metastasis.⁹

The discovery of circulating miRNAs in blood plasma and serum suggests the possibility of using this class of non-coding RNAs as non-invasive biomarkers for an earlier cancer diagnosis and to predict prognosis and response to therapy.¹⁰

By using 2 independent retrospective cohorts of formalin-fixed paraffin-embedded (FFPE) tissue, in a previous work, we identified a group of miRNAs differentially expressed between thymic tumors and normal thymic tissues.¹¹ Among the 14 most significantly deregulated miRNAs (9 up- and 5 down-regulated) we observed the down-regulation of miRNAs with tumor- suppressor function, as miR-145-5p, and the up-regulation of onco-miRNAs, as miR-21-5p and miR-148a-3p and miR-141-3p, in TET, in line with what previously reported for other tumors.^11,12

Based on this evidence, we have collected blood plasma from TET patients at the time of surgery and during follow-up (with 35 months as median of time from surgery), and from healthy donors samples, to evaluate expression levels of the up-regulated miRNAs identified in the tumor. We hereby report that the expression levels of specific onco-miRNAs, that we found upregulated in the blood plasma collected from TET patients at surgery compared to healthy donors samples resulted significantly reduced in follow-up samples. This pilot study suggests that these circulating miRNAs could represent new non-invasive biomarkers for the effectiveness of treatment.

Results

Among the up-regulated miRNAs previously identified in FFPE tumor tissues ¹¹ we select 6 miRNAs (miR-21-5p, miR-148a-3p, miR-141-3p, miR-34b-5p, miR-34c-5p, miR-455-5p) to evaluate their expression in blood plasma and peripheral blood mononuclear cells collected from TET patients at the time of surgery and during follow-up (with 35 months as median of time from surgery), and from healthy donors samples. The characteristics of TET patient specimens are reported in the Patients and Methods section and summarized in Table 1.

Table 1.

Characteristics of Thymoma specimens.

Pts	Sex/age	Histotype/Stage	Tumor size (cm)	MG	WBC/neutrophil x 109/L	Hemoglobin (g/dL)	Platelet x 109/L	Preop. CRT	Completeness of surgical resection	Postop. CT/cycles	Postop. RT/dose	Therapeutic Response	Follow -up
1	M/58	B1 II-a	12 x 10 x 3	Yes	7.12/4.5	14	251	No	R0	Yes 3 cycles	Yes 50 Gy	NED	30 months
2	F/64	B2 I	11 x 7 x 2.5	No	12.99/4.27	12.2	235	No	R0	Yes 3 cycles	Yes 50 Gy	NED	12 months
3	F/32	B2 II-b	8 x 4.8 x 7.5	Yes	6.73/4.04	13	254	No	R0	Yes 3 cycles	Yes 50.4 Gy	NED	48 months
4	F/60	B2/B3 II-a	7 x 3 x 2	Yes	11.4/5.48	10.9	284	No	R0	Yes 3 cycles	Yes 50.4 Gy	NED	44 months
5	M/60	B3 II-b	8 x 4 x 3	No	7.60/4.67	13.8	232	No	R0	Yes 3 cycles	Yes 50.4 Gy	NED	35 months

Although miR-34b-5p, miR-34c-5p, miR-455-5p were found upregulated in the tumor, they resulted undetectable in blood plasma collected from TET patients and healthy donors (data not shown). On the contrary, miR-21-5p and miR-148a-3p resulted significantly up-regulated in blood plasma collected from TET patients at the time of surgery compared to healthy donors samples (Fig. 1A). Of note the expression levels of these onco-miRNAs, in the same TET patients cohort, resulted significantly reduced during follow-up (Fig. 1A). Interestingly, also miR-141-3p expression was modulated similarly to miR-21-5p and miR-148a-3p, although it did not reach statistical significance (Fig. 1A). As internal control we selected miR-940, which resulted unchanged in tumor versus normal FFPE tissue samples in our previous work by array method¹¹ (Fig. S1A). We observed that the modulation of miR-940 expression did not reach statistical significance in blood plasma collected from TET patients at the time of surgery compared to healthy donors samples (Fig. S1B). Moreover, we observed a great degree of variability of expression for miR-940 in the follow-up samples, without reaching any significant correlation (Fig. S1B).

Figure 1.

Evaluation of microRNAs levels in blood plasma and mononuclear cells in TET patients. (A) Box plots showing the modulation miR-21-5p, miR-148a-3p and miR-141-3p in blood plasma collected from 5 TET patients at surgery (Tumor) and during follow-up (Follow-up), and from 5 healthy donors samples (Normal) by RT-qPCR. (B). Box plots showing the modulation miR-21-5p, miR-148a-3p and miR-141-3p in mononuclear cells collected from 5 TET patients at surgery (Tumor) and during follow-up (Follow-up), and from 5 healthy donors samples (Normal), by RT-qPCR.

The expression level of the modulated miRNAs was also evaluated on peripheral blood mononuclear cells collected from the same patients cohort and healthy donors.

We observed a great degree of variability of expression for miR-21-5p, miR-148a-3p and miR-141-3p in the follow-up samples, without reaching any significant correlation (Fig. 1B). Although, miR-21-5p and miR-141-3p expression resulted unchanged in mononuclear cells from TET patients compared to healthy donors, miR-148a-3p expression showed a significant upregulation in tumor vs healthy donor samples (Fig. 1B).

Discussion

In this study we show for the first time that miRNAs that we previously found upregulated in TET vs autologous normal tissues also show increased levels in plasma samples from TET patients when compared with healthy donors. Specifically, among the up-regulated miRNAs we confirmed increased plasma levels for miR-21-5p and miR-148a-3p. MiR-148a-3p resulted significantly up regulated not only in plasma but also in mononuclear cells collected from the same patient cohort. Plasma level of miR-21-5p and miR-148a-3p resulted significantly reduced during follow-up, highlighting these onco-miRNAs as potential biomarker useful not only for diagnosis but also to assess complete resection or response to treatment.

Increased circulating miRNAs levels were previously reported in cancer patients, interestingly, miR-148a was described as a promising predictor biomarker to identify individuals with poor prognosis in patients with osteosarcoma,¹³ miR-21 was indicative of a shorter disease free survival and overall survival in NSCLC patients ¹⁴ and miR-141 was a predictor of poor survival in advanced colon cancer.¹⁵

In summary, this study sheds light on a group of miRNAs that are altered in TET and are emerging as putative non-invasive biomarkers for TET management. These findings might be useful to pave the way for additional and more detailed investigations to clarify the clinical usefulness of circulating miRNAs in TET. In particular, the evaluation of circulating miR-21-5p and miR-148a-3p levels in a larger thymic tumor series which allows comparison of patients with different outcome will define the strength of these miRNAs as prognostic biomarkers.

Patients and methods

Patients

This study was conducted in 5 patients with thymic epithelial tumors undergoing complete surgical resection. Two patients were female, and 3 patients were male. The median patient age was 60 years (range, 32-64). Three patients (60%) were myasthenic. All thymic epithelial tumors were classified according to the 2004 World Health Organization (WHO) histologic classification system and clinical stage determined according to the criteria of Masaoka-Koga.^16,17 One patient was Type B1 (Stage II-A), 2 were Type B2 (Stage I and II-B), 1 Type B2/B3 (Stage II-A), and 1 Type B3 (Stage II-B). No preoperative chemo-radiotherapy was administered (Table 1).

All patients underwent radical thymectomy by median sternotomy. After surgery all patients were referred for adjuvant chemo-radiotherapy with cisplatin (50 mg/m²), adriblastin (50 mg/m²), and cyclophosphamide (500 mg/m²) repeated 3 times every 3 weeks and radiotherapy (RT). Post-operative RT was performed within 3 months from the surgical procedure, for a total dose of 50-50.4 Gy in 1.8-2.0 Gy fractions. Computed tomography scan was performed within 3 months after the end of RT and not evidence of disease (NED) was observed in all the patients (Table 1).

Samples collection

Peripheral blood samples were collected to obtain blood plasma samples and Ficoll-Hypaque-isolated peripheral blood mononuclear cells. Blood plasma samples were obtained by standard venipuncture and centrifugation in EDTA-coated tubes and frozen at -80 °C prior to analysis. The institutional ethical committee board approved the study.

RNA extraction and microRNA expression analysis

Total RNA from 200 μl of plasma samples was extracted using the miRNAeasy serum/plasma kit (Qiagen, Chatsworth, CA) following the manufacturer's instructions. The concentration and purity of total RNA were assessed using a Nanodrop TM 1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA). 50 ng of total RNA was reverse transcribed in 20 μl by using miScript II RT kit (Qiagen, Chatsworth, CA) and 1 μl of cDNA dilution (1:5) was used for quantitative real-time PCR (RT-qPCR) experiments. Quantification of miRNAs (miR-21-5p, miR-148a-3p, miR-141-3p, miR-34b-5p, miR-34c-5p, miR-455-5p) was carried out by miScript Primer Assay (Qiagen, Chatsworth, CA), normalizing over the Spike-In Control (Qiagen, Chatsworth, CA) and using the miScript SYBR Green PCR kit (Qiagen, Chatsworth, CA). All reactions were performed in triplicate.

Total RNA from Ficoll-Hypaque-isolated peripheral blood cells was extracted using TRIzol RNA isolation system (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. 150 ng of total RNA was reverse transcribed in 8 μl of volume using miScript II RT kit (Qiagen, Chatsworth, CA) and 1 μl of the cDNA dilution (1:4) was used for quantitative real-time PCR (RT-qPCR) experiments. ΔΔCt values were normalized with those obtained from the amplification of the endogenous U6B snRNA (Qiagen, Chatsworth, CA). All reactions were performed in triplicate.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

Contribution of AIRC (StG 4841) and Sapienza University of Rome to FF was greatly appreciated.