Kuno Suzuki, Tatsuya Yamaguchi, Masakazu Kohda, Masami Tanaka, Hiroyuki Takemura, Mitsuru Wakita, Yoko Tabe, Shunsuke Kato, Motomi Nasu, Takashi Hashimoto, Shinji Mine, Nobuko Serizawa, Ko Tomishima, Akihito Nagahara, Takahisa Matsuda, Taiki Yamaji, Shoichiro Tsugane, Yutaka Saito, Hiroyuki Daiko, Takaki Yoshikawa, Ken Kato, Takuji Okusaka, Takahiro Ochiya, Yusuke Yamamoto, Shoji Yotsui, Takashi Yamamoto, Tomoyuki Yamasaki, Hiroshi Miyata, Masayoshi Yasui, Takeshi Omori, Kazuyoshi Ohkawa, Kenji Ikezawa, Tasuku Nakabori, Naotoshi Sugimoto, Toshihiro Kudo, Keiichi Yoshida, Masayuki Ohue, and Takashi Nishizawa
The relationship between the expression of microRNAs (miRNAs) in blood and a variety of diseases has been investigated. MiRNA-based liquid biopsy has attracted much attention, and cancer-specific miRNAs have been reported. However, the results of analyses of the expression of these miRNAs vary among studies. The reproduction of results regarding miRNA expression levels could be difficult if there are differences in the data acquisition process. Previous studies have shown that the anticoagulant type used during plasma preparation and sample storage conditions could contribute to differences in measured miRNA levels. Thus, the impact of these preanalytical conditions on comprehensive miRNA expression profiles was examined. First, the miRNA expression profiles of samples obtained from healthy volunteers were analyzed using next-generation sequencing. Based on an analysis of the library concentration, human genome identification rate, ratio of unique sequences and expression profiles, the optimal preanalytical conditions for obtaining highly reproducible miRNA expression profiles were established. The optimal preanalytical conditions were as follows: ethylenediaminetetraacetic acid (EDTA) as the anticoagulant, whole-blood storage at room temperature within 6 hours, and plasma storage at 4°C or -20°C within 30 days. Next, plasma samples were collected from 60 cancer patients (3 facilities × 20 patients/facility), and miRNA expression profiles were analyzed. There were no significant differences in measurements except in the expression of erythrocyte-derived hsa-miR-451a. However, the variation in hsa-miR-451a levels was smaller among facilities than among individuals. This finding suggests that samples obtained from the same facility could show significantly different degrees of hemolysis across individuals. We found that the standardization of anticoagulant use and storage conditions contributed to reducing the variation in sample quality across facilities. The findings from this study could be useful in developing protocols for collecting samples from multiple facilities for cancer screening tests.