Exosomal miR-1290 is a potential biomarker of high-grade serous ovarian carcinoma and can discriminate patients from those with malignancies of other histological types

Dulbecco’s modified Eagle medium and Roswell Park Memorial Institute 1640 medium were obtained from Nacalai Tesque (Kyoto, Japan). Fetal bovine serum (FBS; #172012) was purchased from Sigma Aldrich (St. Louis, MO). Antibody against CD63 (#11–343-C025) was purchased from EXBIO (Vestec, Czech Republic). Donkey anti-mouse IgG 10 nm gold (#ab39593) was purchased from Abcam (Cambridge, UK).

TYK-nu cell line was purchased from Health Science Research Resources Bank (Osaka, Japan). HeyA8 cell line was generously contributed by Dr. Anil Sood (MD Anderson Cancer Center, Houston, TX). IOSE cell line was generously contributed by Dr. Masaki Mandai (Kyoto University, Kyoto, Japan). Briefly, ovarian surface epithelial cells were collected from normal ovary and transfected with SV40 large T antigen and human telomerase reverse transcriptase [13]. These cell lines were cultured in optimal medium according to the suppliers’ recommendations.

Exosomes were isolated from conditioned medium (CM) by an ultracentrifugation method as previously described [14]. The exosome-containing pellet was resuspended in phosphate-buffered saline (PBS) and the amount of exosomal protein was assessed by the Lowry method (Bio-Rad, Hercules, CA). Total RNA was extracted using TRIzol reagent (Life Technologies, Carlsbad, CA: #15596–018).

The morphology and characteristics of exosomes were verified by transmission electron microscopy using a H-7650 transmission electron microscope from Hitachi (Tokyo, Japan) as described [15].

A miRNA microarray using the GeneChip miRNA 4.0 array (Affymetrix, Santa Clara, CA) was performed and analyzed by Filgen (Nagoya, Japan). Briefly, 1000 ng of each miRNA sample was biotin-labeled using a Flash TagTM Biotin HSR RNA labeling kit for Affymetrix GeneChip miRNA arrays (Affymetrix) according to the manufacturer’s protocol. Hybridization solution was prepared using 110.5 μL hybridization master mix and 21.5 μL biotin-labeled sample. The array was incubated using a GeneChip Hybridization Oven 645 (Affymetrix) and washed using a GeneChip Fluidics station 450 (Affymetrix) according to the manufacturer’s protocol. The washed array was analyzed using a GeneChip Scanner 3000 7G (Affymetrix).

In a retrospective cohort study, data were reviewed from patients with ovarian cancer treated at Osaka University Hospital (Suita, Osaka) between January 2013 and May 2015. The study was approved by the Institutional Review Board of the institute and conformed to the provisions of the Declaration of Helsinki. Written informed consent was obtained from every participant for the use of their samples.

Blood samples were collected from healthy volunteers (n = 13) and patients with ovarian cancer (n = 70). From 16 HGSOC patients, blood samples were collected twice at the timing of admission for primary debulking surgery (PDS) (2 or 3 days before surgery) and the initial postoperative chemotherapy (approximately 28 days after surgery). The samples were centrifuged at 1500 x g for 10 min at 4 °C and the upper sera phases were stored in aliquots at − 80 °C until use.

miRNAs were extracted from 200 μL serum samples using an miRNeasy Serum/Plasma Kit (QIAGEN, Hilden, Germany) according to the manufacturer’s protocol. The eluates containing miRNAs were stored at − 80 °C until use.

The expression of miR-1290 was evaluated by quantitative real-time polymerase chain reaction (qRT-PCR). TaqMan miRNA assays (#4366596, Applied Biosystems, Foster City, CA) with specific reverse transcription primers and probes were used to quantify the expression of mature human miR-1290 (5′ – UGGAUUUUUGGAUCAGGGA – 3′). Before RNA extraction, 100 fmol/mL of synthesized cel-miR-39 (#4427975, Applied Biosystems) was added to an equal volume of serum to serve as a normalizer. The relative levels of miR-1290 in serum were expressed using the 2^−ΔCt method, in which ΔCt = Ct^miR-1290 − Ct^cel-miR-39. qRT-PCR was performed in triplicate with the StepOnePlus Real-Time PCR system (Applied Biosystems).

First, by multi-step ultracentrifugation, exosomes were collected from the CM of HeyA8, TYK-nu and IOSE. HeyA8 and TYK-nu were derived from HGSOC cells and IOSE is an immortalized normal ovarian epithelial cell line used as a benign control. Exosomes exhibited a round-shaped morphology, showed positive staining for CD63, which is a representative exosomal marker, and were approximately 100 nm in size as determined by electron microscopy (Fig. 1a) [16]. miRNA microarray analysis revealed the expression profiles of miRNAs which were highly expressed in exosomes derived from the HGSOC cells compared with those from the IOSE cells (Fig. 1b). Among those, 4 miRNAs (miR-129b-1-3p, miR-139-5p, miR-1290, and miR-3131) were identified as highly expressed (> 4.0) in both HeyA8- and TYK-nu-derived exosomes (Fig. 1c). miRNA qRT-PCR assay validated that the expression level of miR-1290 was higher by 5.41-fold in HeyA8 exosomes and 8.92-fold in TYK-nu exosomes compared with that in IOSE exosomes (Fig. 1d).

Since the expression of miR-1290 was highly expressed in exosomes derived from HGSOC cell lines, we were encouraged to determine miR-1290 expression levels in the sera of patients with EOC and to evaluate its diagnostic value as a biomarker. A total of 83 sera samples were collected from healthy controls (n = 13) and patients with EOC (n = 70), and miR-1290 expression levels were examined by qRT-PCR. Clinical characteristics of patients and healthy controls are presented in Table 1. Out of 70 EOC patients, 30 (43%) had serous ovarian cancer, all of which were diagnosed as HGSOC. Among these, 22 (72%) had advanced stage (stage III/IV) disease and 8 (27%) had early stage (stage I/II) disease (Table 2).

The relative expression level of miR-1290 in EOC patients was 2.26 compared to that in healthy controls (Fig. 2a); however, there was no significant difference between the 2 groups (P = 0.89). In contrast, CA125 value in EOC patients was 311.34 U/mL and significantly elevated comapred to that in healthy controls (11.99 U/mL) (P < 0.01; Fig. 2b). The diagnostic performance value of serum miR-1290 was calculated by ROC curve analysis to discriminate EOC patients from healthy controls. When the cut-off of miR-1290 was set at 1.49, the sensitivity and specificity were 0.51 and 0.57, respectively (AUC = 0.48; Fig. 2c). For comparison, a ROC curve to discriminate EOC patients from healthy controls was similarly calculated for CA125. When the cut-off of CA125 was set at 76 U/mL, the sensitivity and specificity were 1.00 and 0.61, respectively (AUC = 0.90). Thus, CA125 served as a better biomarker to disciminate all EOC patients from healthy controls. The combination of the expression levels of miR-1290 and CA125 for ROC curve analysis did not significantly improve the AUC value (from 0.90 to 0.92; P = 0.08; Fig. 2c), indicating that this model failed to distinguish EOC patients from healthy controls. EOC consists of the following 4 major histological subtypes: serous (in our case series, all such cases were diagnosed as HGSOC), clear-cell, endometrioid, and mucinous. Using a cut-off of 1.20, the miR-1290 classification model had AUC, sensitivity, and specificity values of 0.71, 0.63, and 0.85, respectively, for HGSOC; 0.69, 0.58, and 0.89, respectively, for clear-cell EOC; 0.62, 0.50, and 0.83, respectively, for endometrioid EOC; and 0.78, 0.58, and 0.90, respectively, for mucinous EOC (Fig. 2d). Among these, the diagnostic performance of HGSOC was good and the combination of miR-1290 and CA125 significantly improved AUC value from 0.71 to 0.97 (P < 0.01; Fig. 2d). Indeed, the relative miR-1290 expression in HGSOC patients was 3.52 and significantly higher than that in healthy controls (P = 0.03; Fig. 2e). Collectively, our data suggested that serum miR-1290 might be a useful biomarker to detect HGSOC in patients.

While the relative expression of miR-1290 in stage I-II EOC patients was 1.58, that in stage III-IV patients was increased to 4.23 (P = 0.23; Fig. 3a). In order to find out whether miR-1290 reflected tumor burden, sera were collected from 16 HGSOC patients twice just before PDS and approximately 28 days after PDS as well as before the following chemotherapy, and its expression level was compared (Fig. 3b). While the relative expression level before PDS was 5.87, that after PDS significantly declined to 1.17 (P < 0.01), suggesting that serum miR-1290 reflects tumor burden and implying that it is derived from HGSOC cells.

Several phase III studies have shown that interval debulking surgery following neoadjuvant chemotherapy (NAC) is less invasive with non-inferior clinical outcomes compared to PDS followed by chemotherapy for treating advanced stage ovarian cancer [17‐19]. However, Tian et al. showed that platinum-based first-line chemotherapy was much less effective for mucinous and clear-cell types, although it is effective for serous and endometrioid types [20]. Therefore, it would be clinically beneficial if we could establish an appropriate way to determine which patients should be treat with NAC. For this reason, it is meaningful to find a biomarker specific to HGSOC which is able to discriminate it from EOC of other histological types. Among 70 sera samples from EOC patients, the expression level of miR-1290 was compared between HGSOC patients (n = 30) and patients with EOC other than HGSOC (n = 40). While the expression level of miR-1290 in patients with EOC other than HGSOC was 1.32, that in HGSOC patients was 3.52, which was significantly higher (P < 0.01, Fig. 4a). In contrast, while CA125 value in patients with EOC other than HGSOC was 239.12 U/mL, that in HGSOC patients was 505.73 U/mL and no significant differences were seen (P = 0.08; Fig. 4b). When the cut-off of the relative miR-1290 expression was set at 1.55, the sensitivity was 0.47 and the specificity was 0.85 (AUC = 0.76; Fig. 4c). For comparison with CA125, ROC curve analysis for discriminating patients with HGSOC from those with EOC other than HGSOC was similarly calculated. The sensitivity and specificity of serum CA125 were 0.73 and 0.63, respectively (AUC = 0.69). Although there was no significant difference between the AUC values of miR-1290 and CA125, the combined AUC value was 0.79 and improved from the AUC values of either alone.