Informed consent was obtained from CRC patients and non-cancer patients for the use of their blood samples. Venous blood samples were collected from the CRC patients (n=114) and non-cancer patients (n=32) from April 2011 to June 2013. In 30 of the CRC patients, blood samples were taken before and 7th day after surgery. No cancer patients received chemotherapy or radiotherapy before blood sampling. The blood samples were obtained from Osaka University, and its associated hospitals. Whole blood was collected, centrifuged at 1,000 rpm at 4°C for 15 min. The supernatant fluids were centrifuged at 15,000 rpm at 4°C for 10 min. The supernatant fluids were stored at −80°C until RNA extraction. This study was conducted under the supervision of the ethics board of Osaka University Hospital.

Small RNA was enriched from all serum samples using the mirVana PARIS RNA isolation kit (Ambion, Austin, TX, USA), following the manufacturer's instructions. Briefly, 400 μl of serum was thawed on ice and centrifuged at 15,000 rpm for 15 min to remove cell debris. Next, 300 μl of supernatant was lysed with an equal volume of 2X denaturing solution. For normalization of sample-to-sample variation during the RNA isolation procedures, 20 fmol of synthetic C. elegans miRNA cel-miR-39 was added to each denatured sample. Small RNAs were then enriched and purified following manufacturer's protocol. The concentration of all RNA samples were quantified by Nano Drop ND-1000 (NanoDrop Technologies, Wilmington, DE, USA).

miRNA microarray experiments were carried out by using Agilent human miRNA microarray cataloged in the Sanger database version 12.0 (design ID 021827). About ten nanogram aliquots of total RNA with cel-miR-39 was used for making miRNA probes according to the Agilent protocol (version 2.3). Microarrays were performed for paired pre- and post-operative serum from 10 CRC patients. Briefly, total RNA was dephosphorylated with calf intestine alkaline phosphatase, denatured with dimethyl sulfoxide, and labeled with pCp-Cy3 using T4 RNA ligase using the miRNA Labeling Reagent and Hybridization kit. Probes were hybridized at 55°C for 20 h with rotation. Then the slides were washed by Gene Expression Wash Buffer 1 at room temperature for 5 min and by Gene Expression Wash Buffer 2 at 37°C for 5 min. After hybridization and washing, the slides were scanned using an Agilent scanner (G2505C). Images were extracted using Agilent Feature Extraction software (version 10.7.3.1) and Agilent GeneSpring GX software (version 10.0.2). Differences in miRNA expression between the 10 pairs was determined if the fold change of cel-miR-39 normalized expression values was >2.0 and the P-value was <0.05 using paired t-test for further analysis. The microarray raw data are available in Gene Expression Omnibus (GEO;http://www.ncbi.nlm.nih.gov/geo) under accession number GSE55139.

For microRNA based RT-PCR assays, 2.5 μl of enriched small RNAs from serum samples were reverse transcribed using the TaqMan MicroRNA Reverse Transcription kit (Applied Biosystems, San Diego, CA, USA) according to manufacturer's instructions in a total reaction volume of 7.5 μl. A 1:20 dilution of RT products was used as template for the PCR stage. PCR reaction was performed in triplicates using TaqMan 2X Universal PCR Master Mix according to the manufacturer's instructions. Each reaction was performed in a final column of 20 μl containing 1.33 μl of the cDNA and 1 μl of Taqman miRNA assay mix. The amplification profile was: denaturation at 95°C for 10 min, followed by 40 cycles of 95°C for 15 sec and 60°C for 60 sec. Each sample was run in triplicates for analysis. The cycle threshold (Ct) is defined as the number of cycles required for the fluorescent signal to cross the threshold in qPCR. The 7900 Sequence Detection System 2.3 (Applied Biosystems) software was used to compute the relative change in RNA expression by the 2−ΔΔCt method with 95% confidence intervals.

The miRNA-specific primer sequences, including miRNA-103, miR-720, cel-miR-39, and RNU6B were designed based on the miRNA sequences obtained from the miRBase. The primer sequences are: hsa-miR-103, 5′-AGCAGCAUUGUACAGGGCUAUGA-3′; hsa-miR-720, 5′-UCUCGCUGGGGCCUCCA-3′; hsa-miR-21, 5′-UAGCUUAUCAGACUGAUGUUGA-3′; cel-miR-39, 5′-UCACCGGGUGUAAAUCAGCUUG-3′; RNU6B, 5′-CGCAAGGATGACACGCAAATTCGTGAAGCGTTCCATATTTTT-3′.

The significance of serum miRNA level was determined by Mann-Whitney test, Wilcoxon test and χ² test where appropriate using the Graph Pad Prism 6 (San Diego, CA, USA). The sensitivity, specificity, and accuracy were calculated according to the standard formulas. Receiver operating characteristic (ROC) curves and area under the ROC curve (AUC) were established for discriminating patients with CRC. P-value of <0.05 was considered statistically significant.

In comparison of CRC patient serum between pre- and post-operation (n=10; 4 stage II CRCs and 6 stage III CRCs) by the miRNA array, we identified miRNAs, most of which showed a decrease in the post-operative serum (Fig. 1). Among them, we focused on the two miRNAs, miR-103 and miR-720 whose expression showed a large decrease after surgery with significant P-value (4.09-fold and 4.68-fold decrease; P=0.037 and P=0.037, respectively, Fig. 1).

We then attempted to confirm the results obtained by miRNA array by qRT-PCR in extended samples of CRC patients (n=30). As shown in Fig. 2, significant decrease in miRNA levels was noted in the post-operative serum for both miRNAs; P=0.0004 for miR-103, and P=0.027 for miR-720.

We examined serum miR-103 and miR-720 levels in 32 non-cancer patients and 84 CRC patients. As shown in Fig. 3, both miRNAs were significantly higher in CRC patients than non-cancer patients (P=0.0067 for miR-103, and P=0.030 for miR-720, respectively). ROC curve was drawn for serum miR-103 and miR-720, which yielded 0.662 for miR-103 and 0.630 for miR-720 as a value of AUC (Fig. 3). In discriminating CRC from non-tumor control subjects, the sensitivity and specificity of miR-103 were 55.9% and 75.0% at a cut-off point of 0.00081; these values for miR-720 were 58.3% and 56.3% at a cut-off point of 0.0016.

As a reference we examined miR-21 as a putative circulating miRNA. Using the same serum sets of normal and CRC patients, we found that serum miR-21 levels increased significantly in CRC patients as compared to non-tumor control patients (P=0.003). AUC of the ROC curves was 0.675. When a cut-off point was set at 0.0107, the sensitivity was 54.7% and the specificity was 84.4% in discriminating CRC from non-tumor control subjects (Fig. 4).

CRC patients were divided into two groups by a median expression. Clinical and pathological survey indicated that high expression of miR-103 was significantly associated with poor differentiation and lymphatic invasion as compared to low miR-103 expression group (P=0.044 and P=0.040, respectively; Table I). High expression of miR-720 was significantly associated with male gender and lymph node metastasis as compared to low miR-720 expression group (P=0.001 and P=0.048, respectively; Table I).

We examined miR-103 and miR-720 expression in normal colonic mucosa (n=10) and CRC tissue samples (n=10). Both miRs were significantly lower in CRC tissue samples than in normal colonic mucosa (P=0.034 for miR-103, and P=0.0052 for miR-720) (Fig. 5).