Fecal miR-106a Is a Useful Marker for Colorectal Cancer Patients with False-Negative Results in Immunochemical Fecal Occult Blood Test

Cancer screening is important for various cancers, including colorectal cancer, to reduce the mortality rates. Fecal occult blood test (FOBT) has been widely used as a screening test for colorectal cancer (1–3) and is classified into two different methods: one is chemical FOBT, such as guaiac FOBT (gFOBT), and the other is immunochemical FOBT (iFOBT). The sensitivity of iFOBT to detect colorectal cancer and advanced adenoma is higher than that of gFOBT (4, 5). Several studies have investigated the efficacy of FOBT using total colonoscopy (TCS) as a reference standard in all participants (6–9). Three studies were conducted on the basis of gFOBT and the other was on iFOBT. However, these large-scale studies showed that the sensitivity of FOBT is not very high. The problem with FOBT is that some colorectal cancer patients present with false-negative results whereas healthy individuals have false-positive ones.

Several attempts to use molecular biological methods for the early detection of colorectal cancer have been reported. In fecal DNA-based analysis, the stool DNA test (sDNA test; ref. 8) was recommended as one approach (10). We also reported several DNA-based methods for the detection of early colorectal cancer using direct sequence analysis (11) and single-strand conformation polymorphism analysis (12) in exfoliated colonocytes. However, the sensitivity and specificity of the sDNA test were insufficient compared with those of gFOBT (13). In fecal RNA-based analysis, several attempts to detect colorectal cancer by using reverse transcription PCR (RT-PCR) on fecal samples have been reported (14–16). However, no molecular biological method has proved to be superior to iFOBT in terms of sensitivity and specificity.

MicroRNAs (miRNA) are small (18–25 nucleotides in size) noncoding RNA molecules that regulate the activity of specific mRNA targets and play various roles in cancer. Several recent studies have clarified that circulating miRNA in plasma is a potential marker for colorectal cancer detection (17, 18), and is remarkably stable in plasma due to its protection from endogenous RNase activity (19). miRNAs can be preserved in poor conditions, such as formalin-fixed paraffin-embedded sections stored for 20 years (20) and fecal samples (21). Thus, miRNAs could be worthy of investigation as tumor biomarkers. Recently, we have reported that miRNA could be extracted from the residuum of iFOBT at a level sufficient for analysis and that fecal miRNA could become a tool for the detection of colorectal cancer (22).

In this study, fecal miRNA extracted from the residuum of iFOBT for colorectal cancer patients and healthy volunteers with no abnormal lesions was assessed as a fecal miRNA test (FmiRT) to determine whether the combination of FmiRT with iFOBT can improve the rate of false-negative results compared with those obtained using iFOBT alone in colorectal cancer screening.

From October 2009 to September 2012 and from October 2010 to December 2011, 117 patients with colorectal cancer and 107 healthy volunteers were enrolled in this study. Their characteristics are summarized in Table 1. All the patients had undergone surgical resection of their primary cancer at the National Cancer Center Hospital, Tokyo, Japan, or the National Cancer Center Hospital East, Kashiwa, Japan. The median age of the patients was 65 years. The locations of the primary tumor were right colon in 32 patients (27.4%) and left colon in 85 (72.6%). The median diameter of the primary tumor was 33 mm. The clinical stages of the patients were early stage (Dukes stage A and B) in 76 (65.0%) and advanced stage (Dukes stage C and D) in 41 (35.0%). All the healthy volunteers were shown to have no symptoms or evident abnormalities, such as adenoma or carcinoma (including hyperplastic polyps), by screening colonoscopy conducted at the Research Center for Cancer Prevention and Screening, National Cancer Center, Tokyo, Japan. The median age of the volunteers was 60 years. All participants were provided with detailed information about the study, and each gave written consent to participate in it, the protocol of which was approved by the Institutional Review Board of the National Cancer Center, Japan.

Naturally evacuated fecal samples were obtained from colorectal cancer patients before they underwent surgical resection. Fecal samples were also obtained from healthy volunteers a few weeks after they had undergone screening colonoscopy. All participants were instructed to evacuate at home into a disposable 5 cm × 10 cm polystyrene tray (AsOne) and bring the sample to the outpatient clinic or the Cancer Prevention and Screening Center of the National Cancer Center, Japan. The fecal samples were prepared for the next step immediately after they were brought to our laboratory, and excess feces was stored at −80°C.

Ten-milligram fecal samples were transferred into the fecal sampling container provided with the immunochemical fecal occult blood test (iFOBT) kit, OC-Hemocatch (Eiken Chemical). iFOBT was immediately conducted using OC-Hemocatch in accordance with the manufacturer's instructions. Briefly, 100-μL samples of the dissolved feces were incubated with the iFOBT kit reagents for 5 minutes at room temperature. After incubation, a blue line seemed in the control window if the iFOBT had been conducted correctly. iFOBT was designated as positive if a blue line also seemed in the test window, whereas the samples for which a line did not seem in the test window were designated as negative. The cut-off value for the concentration of hemoglobin in the iFOBT kit was 50 ng/mL.

The solution remaining after iFOBT analysis was collected into a 2-mL tube and centrifuged at 700 × g for 5 minutes at 4°C, following which the fecal pellet was collected by removal of the supernatant. Total RNA was extracted using an miRNeasy Mini Kit (Qiagen) according to the manufacturer's instructions. Briefly, each fecal pellet was resuspended in 1 mL of QIAzol (Qiagen). Thereafter, 200 μL of chloroform was added, the mixture was vortexed vigorously for 15 seconds and incubated for 3 minutes at room temperature, and then centrifuged at 13,000 × g for 15 minutes at 4°C. The upper aqueous phase was transferred to a 1.5-mL tube, and 1.5 volumes of 100% ethanol were added. The solution was mixed thoroughly by pipetting and transferred to an miRNeasy spin column, and the columns were centrifuged at 7,000 × g for 15 seconds at room temperature. After washing, total RNA was dissolved in 100 μL of RNase-free water.

Total RNA extracted directly from feces was also obtained. Fecal samples were homogenized as described previously (21), and total RNA was extracted using an miRNeasy Mini Kit by almost the same procedure as described earlier. Briefly, 500 mg of feces was homogenized with 5 mL of QIAzol using an IKA Ultra-Turrax homogenizer (IKA-Werke) at 6,000 rpm for 2 minutes at room temperature. The homogenates were centrifuged at 13,000 × g for 5 minutes at 4°C. The supernatants were then transferred to a new tube, and up to 5 mL more of QIAzol was added, followed by 1.5 mL of chloroform. The tubes were shaken vigorously for 30 seconds and centrifuged at 13,000 × g for 15 minutes at 4°C. The aqueous phase was then transferred into a new tube. One and a half volumes of 100% ethanol were added, and the solution was mixed thoroughly by pipetting. The mixture was then poured onto an miRNeasy spin column and the column was centrifuged at 7,000 × g for 15 seconds at room temperature. After washing, total RNA was dissolved in 100 μL of RNase-free water.

The concentrations of total RNA were determined using a NanoDrop UV spectrometer (LMS). The RNA samples were stored at −80°C until analysis. Total RNA extracted from residuum of iFOBT is referred to as pellet RNA and total RNA directly extracted from feces was named as direct RNA.

cDNA was synthesized using the TaqMan MicroRNA RT Kit (Applied Biosystems), in accordance with the manufacturer's instructions. The reaction mixture consisted of 5 ng of total RNA, 0.5 μL of 10 × RT buffer, 1 μL of 5× specific primer, 0.05 μL of dNTPs (100 mmol/L), 0.06 μL of RNase inhibitor (20 U/μL), and 0.33 μL of MultiScribe Reverse Transcriptase (50 U/μL) in a final reaction volume of 5 μL. cDNA synthesis was conducted with incubation at 16°C for 30 minutes and 42°C for 30 minutes.

The reaction mixture for the real-time PCR analysis consisted of 4 μL of template cDNA, 10 μL of TaqMan Fast Universal PCR Master Mix (Applied Biosystems), and 1 μL of 20 × primer/probe mixture in a total reaction volume of 20 μL. Real-time PCR was conducted with precycling heat activation at 95°C for 20 seconds, followed by 40 cycles of denaturation at 95°C for 3 seconds and annealing/extension at 60°C for 30 seconds, in an Applied Biosystems 7500 Fast Real-Time PCR System. For miRNA expression analysis, we targeted 14 miRNAs: miR-15b, -16, -19a, -19b, -20b, -92a, -93, -106a, -132, -142-3p, -223, -223*, -342-3p, and -451, and miR-24 as an internal control. For all of these miRNAs, we used the commercially available TaqMan MicroRNA Assay (Applied Biosystems).

The correlations between “direct miRNA” and “pellet miRNA” were analyzed using Pearson's product–moment correlation coefficient. The miRNA expression analysis was conducted using the comparative Ct (threshold cycle) method (available at http://pathmicro.med.sc.edu/pcr/realtime-home.htm). In this analysis, the formulae for the relative quantification of each of the miRNAs were as follows: (dCt of each miRNA) = (Ct of each miRNA) − (Ct of miR-24), and (relative quantification of each miRNA) = 2^{−(dCt of each miRNA)}. Differences in Ct values of the target miRNAs were analyzed by two-sided Mann–Whitney U test and differences in relative quantification of the target miRNAs between 2 subgroups of colorectal cancer patients and healthy volunteers were analyzed by Games–Howell test of one-way ANOVA. Statistical analyses were conducted using SPSS Statistics Ver. 19 (IBM). P < 0.05 was considered statistically significant.

The median amounts of pellet RNA (from 10 mg of feces) and direct RNA (from 500 mg of feces) in colorectal cancer patients were 1.96 μg (range, 0.33–5.64) and 53.35 μg (range, 6.63–243.72), respectively (Table 1). Those in healthy volunteers were also 1.59 μg (range, 0.33–6.00) and 43.04 μg (range, 6.18–163.79), respectively. The amount of direct RNA was dozens of times higher than that of pellet RNA because of a greater starting volume of fecal sample. The correlations between pellet RNA and direct RNA are shown in Fig. 1. Significant positive correlations were observed between direct RNA and pellet RNA in terms of miR-92a (P < 0.001, r = 0.549), miR-106a (P < 0.001, r = 0.747), and miR-24 (P < 0.001, r = 0.684). These results show that pellet RNA could be applied to FmiRT as effectively as direct RNA.

The Ct values of miR-92a, -106a, and -24 in all participants are shown in the box and whisker plots in Fig. 2. Their median values in pellet RNA from colorectal cancer patients were 29.65 (range, 22.65–40), 29.99 (22.18–35.95), and 28.31 (21.55–33.26); however, those in pellet RNA from healthy volunteers were 30.24 (25.99–40), 30.70 (27.60–33.61), and 28.32 (25.31–31.41). Between the colorectal cancer patients and healthy volunteers, significant differences were observed in the Ct values of miR-92a (P = 0.016) and miR-106a (P < 0.001); however, there was no significant difference in the Ct values of miR-24 (P = 0.67).

On the basis of the above results, miR-24 was used as an internal miRNA control in this study, as in our previous work (22, 23). The median relative expression levels of miR-15b, -16, -19a, -19b, -20b, -92a, -93, -106a, -132, -142-3p, -223, -223*, -342-3p, and -451 in the colorectal cancer patients and the healthy volunteers were shown in Table 2. Significant differences were observed regarding the relative expression levels of miR-16, -19a, -92a, -106a, -132, -142-3p, -223, -223*, and -451 between the colorectal cancer patients and healthy volunteers (P < 0.05). Moreover, particularly significant differences were found for miR-16, -92a, -106a, -142-3p, -223, and -451 (P < 0.01).

To improve the relatively low sensitivity of iFOBT, it is important to identify colorectal cancer patients with false-negative results correctly as positive. In this study, the numbers of true-positive colorectal cancer patients and false-negative colorectal cancer patients using iFOBT were 71 (60.7%) and 46 (39.3%), respectively (Table 1). The median tumor sizes of colorectal cancer patients with true-positive iFOBT (true-positive subgroup) and false-negative iFOBT (false-negative subgroup) were 37 mm (range, 13–125 mm) and 21 mm (7–82mm), respectively. There were significant differences between the true-positive subgroup and the false-negative subgroup regarding the tumor size (P < 0.001) and tumor depth (P = 0.011). However, no significant differences between the true-positive subgroup and the false-negative subgroup were observed for the tumor location and cancer stage (P > 0.05). The relative expression levels of 6 selected miRNAs (miR-16, -92a, -106a, -142-3p, -223, and -451) in the true-positive subgroup, the false-negative subgroup, and the healthy volunteer group are shown in Fig. 3. The expression of miR-451 in the true-positive subgroup was not significantly different from that in the healthy volunteer group (P = 0.122). However, significant differences were observed in the expression of miR-16, -92a, -106a, -142-3p, and -223 between the true-positive subgroup and the healthy volunteer group (P < 0.05). A significant difference was also observed for only miR-106a between the false-negative subgroup and the healthy volunteer group (P < 0.001); however, no significant difference was observed between the false-negative subgroup and the true-positive subgroup (P = 0.291). Thus, miR-106a was used for FmiRT because its expression may reflect the presence of colorectal cancer cells, but not intestinal bleeding into feces.

The sensitivities of FmiRT using miR-106a expression in the true-positive subgroup and the false-negative subgroup were 39.4% [28 of 71; 95% confidence interval (CI), 28.0–51.7] and 26.1% (12 of 46; 95% CI, 14.3–41.1%) using a threshold of miR-106a calculated by volunteer's data that 97% of volunteers were diagnosed as negative result of FmiRT. The sensitivity and specificity of the method combining iFOBT and FmiRT were 70.9% (83 of 117; 95% CI, 61.8–79.0) and 96.3% (103 of 107; 95% CI, 90.7–99.0), respectively (Table 3). One quarter of colorectal cancer patients (12 patients) with false-negative results of iFOBT seemed as true positive by adding FmiRT using fecal miR-106a.

We previously reported the analysis of fecal miRNA using exfoliated colonocytes isolated from feces (23, 24) and several studies have reported that fecal miRNA could be useful for the diagnosis of colorectal cancer using RNA directly extracted from feces (25–27) because miRNA extracted directly from feces was more stable than that from the residuum of iFOBT. However, the sensitivity and specificity of an FmiRT are inferior to those of iFOBT in the several reports. We thought that the issue of iFOBT (the presence of false-negative colorectal cancer patients) could be resolved by combination of iFOBT and FmiRT, thus we used the residuum of iFOBT because “pellet miRNA” is more comparable than “direct miRNA.” Very recently, we also reported that miRNA could be extracted at a level sufficiently for analysis from the residuum of iFOBT (22); however, there have been no studies comparing the miRNA expression between direct fecal RNA and pellet fecal RNA. In reality, we could show the positive correlation between “direct miRNA” and “pellet miRNA” in this study. miRNAs capsuled in exosome could be detected in the supernatant of fecal solution or cell culture as we reported previously (21). However, the fecal pellet contains exfoliated cancer cells, and we thought that colorectal cancer-related miRNAs could be extracted from fecal pellet more efficiently. In addition, our preliminary study revealed that PCR using 100 or 500 ng of fecal miRNA was not succeeded because of the presence of PCR inhibitory molecules contained in feces, and 5 ng of miRNA was enough for analysis of miRNA expression. Because FOBT is well known to produce a large number of false-positive results, Calistri and colleagues assessed whether the long DNA test could improve its specificity in a large series of consecutively enrolled iFOBT+ individuals by reducing false-positive results using residuum of iFOBT (28). They concluded that iFOBT and long DNA analysis were largely independent variables, and that an approach combining iFOBT and long DNA evaluation could help to reduce the rate of unnecessary TCS for false-positive healthy individuals. Thus, the numbers of false-negative colorectal cancer patients and false-positive healthy individuals might be reduced by applying molecular biological methods using fecal DNA or RNA extracted from the residuum of iFOBT.

In our preliminary study, the expressions of fecal miRNA (direct miRNA) from 10 patients of colorectal cancer and 10 healthy volunteers were analyzed using TaqMan Array Human MicroRNA Cards Set v3.0 (Applied Biosystems). A total of 22 miRNAs of 749 miRNAs were selected using following criteria: (i) the fecal miRNA expression in colorectal cancer patients was 2 times higher than that in healthy volunteers, (ii) the mean Ct value of each miRNA was less than 36, and (iii) miRNA was detected in 50% of patients and volunteers. Custom array card contained with these 22 miRNAs (Applied Biosystems) were obtained, and 14 miRNAs of 22 miRNAs were expressed significantly high in colorectal cancer patients compared to healthy volunteers. Thus, 14 miRNAs were used in this study.

The large numbers of false-negative colorectal cancer patients and false-positive healthy individuals in FOBT have long been a concern. In this study, we focused on reducing the size of the former group. It is important in colorectal cancer patients with false-negative results in iFOBT for this error to be corrected because a correct early diagnosis greatly increases the likelihood of curative treatment. In this study, no significant difference between the true-positive subgroup and the false-negative subgroup was observed regarding the expression of miR-106a in colorectal cancer patients. However, there was a significant difference for this marker between the false-negative subgroup in colorectal cancer patients and the healthy volunteer group. However, no significant differences were observed between the false-negative subgroup and the healthy volunteer group in terms of the expression of miR-16, miR-142-3p, and miR-223, whereas there were significant differences between the true-positive and the false-negative subgroups. These results indicate that miR-16, miR-142-3p, and miR-223 may be diagnostic markers of intestinal bleeding, as reflected by the presence of factors such as granulocytes in the feces.

miR-106a is known to be expressed high in colorectal cancer tissue than in normal colorectal tissue, and to suppress several gene expressions. Recently, the expression of TGFBR2 could be suppressed by miR-106a resulting in the enhancement of the cancer invasion (29). Thus, we consider that miR-106a is a suitable marker for cancer screening. The usefulness of fecal miR-106a to detect colorectal cancer has been reported recently (25). However, it is difficult to diagnose the colorectal cancer using fecal miRNA alone. Thus, we conducted to combine iFOBT and fecal miR-106a. The fecal miRNA assay using residuum of iFOBT to detect colorectal cancer has been reported recently (22). The solution of iFOBT is consisted of PBS or HEPES buffer, thus, the exfoliated colorectal cancer cells are stable in the solution at 4°C up to 5 days. In this study, one quarter of colorectal cancer patients (12 patients) with false-negative results of iFOBT seemed as true positive using fecal miR-106a. We think that this is a novel finding. Thus, fecal miR-106a was shown to be a good molecular marker to identify colorectal cancer patients from among iFOBT-negative participants. Combinatory use of FmiRT and iFOBT may thus enhance the sensitivity to detect colorectal cancer.

No potential conflicts of interest were disclosed.

Conception and design: Y. Koga, Y. Matsumura

Development of methodology: Y. Koga, Y. Matsumura

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Koga, N. Yamazaki, Y. Yamamoto

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Koga, N. Yamazaki, Y. Yamamoto, Y. Matsumura

Writing, review, and/or revision of the manuscript: Y. Koga, Y. Matsumura

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): S. Yamamoto, N. Saito, Y. Kakugawa, Y. Otake, M. Matsumoto

Study supervision: Y. Matsumura

The authors thank Ms. N.F. Abe and Ms. M. Ohmaru for their technical assistance and Ms. K. Shiina for her secretarial assistance.

This work was supported by a Grant-in-Aid for the Program for Promotion of Fundamental Studies in Health Sciences of the National Institute of Biomedical Innovation (NIBIO) of Japan (Y. Koga); the Innovation Promotion Program from the New Energy and Industrial Technology Development Organization (NEDO) of Japan (Y. Matsumura); the Japan Society for the Promotion of Science (JSPS) through the Funding Program for World-Leading Innovative R&D on Science and Technology (FIRST Program), initiated by the Council for Science and Technology Policy (CSTP; Y. Matsumura); and the National Cancer Center Research and Development Fund (Y. Matsumura).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.