A Distinct Expression Pattern of the Long 3′-Untranslated Region Dicer mRNA and Its Implications for Posttranscriptional Regulation in Colorectal Cancer

MicroRNAs (miRNAs) are small endogenous noncoding RNAs that act as negative regulators of gene expression.1 Primary miRNAs are transcribed by RNA polymerase II and are processed by RNase III Drosha into long primary miRNAs in the nucleus. The primary miRNAs are transported to the cytoplasm, where they are processed by RNase III Dicer into mature miRNAs, which are able to bind to the 3′-untranslated region (UTR) of target messenger RNAs (mRNAs). The bound miRNAs mediate either the degradation of the mRNA or inhibition of its translation.2 It has been observed that miRNAs can act as oncogenes (oncomiR) or tumor suppressor genes (tsmiR).3 In fact, the downregulation of tsmiR expression is a general trait of human cancers.4 Several lines of evidence have shown that epigenetic silencing and decreased expression of Dicer are implicated in the reduction of tsmiR expression.5, 6 Colorectal cancer (CRC) is one of the most common causes of cancer-related mortality worldwide. The expression of number of miRNAs is dysregulated in CRC; let-7a and miR-101 are decreased,7, 8 whereas miR-21 and miR-92 are increased.9, 10 Recent evidence suggests that altered expression of Dicer is associated with dysregulated miRNA expression in various malignancies including breast and ovarian cancer as well as CRC.11, 12, 13 The regulation of Dicer expression remains unclear but several recent studies have demonstrated that Dicer expression is regulated by miR103/107 and let-7 miRNA.14, 15, 16 Cleavage and polyadenylation are essentially universal steps in the maturation of eukaryotic mRNA transcripts.17, 18 The pre-mRNA is cleaved downstream of the polyadenylation signal (typically 5′-AAUAAA-3′) and polyadenylation occurs at the 3′ end of the mRNA. Over half of all mammalian genes including Dicer have one or more polyadenylation sites,19, 20 resulting in mRNA isoforms that encode the same protein but have 3′UTRs of different lengths. The Dicer mRNA (RefSeq NM_030621) 3′UTR is ~4 kb in length and has two polyadenylation signals, which results in mRNAs of different lengths. One Dicer mRNA is ~6 kb and uses the upstream polyadenylation signal (short 3′UTR Dicer mRNA), whereas another is ~10 kb, and uses the downstream polyadenylation signal (long 3′UTR Dicer mRNA). If more than one polyadenylation signal occurs in a transcript, the dominant signal is usually located downstream. Polyadenylation site usage can be regulated by physiological conditions such as cell growth and differentiation or by pathological events such as cell transformation.15, 21, 22 miRNA-binding sequences are usually located in the 3′UTR of transcripts, so that alternative polyadenylation site usage can affect gene expression through posttranscriptional regulation.23 In fact, the Dicer mRNA with a short 3′UTR has increased mRNA stability and higher protein expression.15 Several studies have been made on Dicer expression and its posttranscriptional regulation by miRNAs in different cancers, but little is known about Dicer alternative polyadenylation site usage in CRC. To investigate the alternative polyadenylation site usage by Dicer, we examined the expression of total Dicer mRNA, the long 3′UTR Dicer mRNA, and miR-103 in CRC. Small RNAs were diluted 1,000-fold with nuclease-free water and a 3-μl aliquot of the diluted RNA was reverse transcribed by Multiscribed Reverse Transcriptase and miRNA-specific primers (Life Technologies). Quantitative PCR for miR-103 and RNU6B (assays ID 000439 and 001093, respectively; Life Technologies) was performed with TaqMan MicroRNAs Assays using a Applied Biosystems 7500 Fast Real-Time PCR system. RNU6B was used as an internal control. A standard reference curve was established for each marker using serial 10-fold dilutions of the recombinant plasmid DNA containing the target sequence or cDNA synthesized from synthetic target RNA. Each sample was run in triplicate and a negative control without template was also run in each reaction plate. The data presented in this study show that: (i) total Dicer mRNA expression in CRC is lower than in adjacent normal mucosa, (ii) long 3′UTR Dicer mRNA expression in CRC is similar to adjacent normal mucosa, and the ratio of long 3′UTR Dicer mRNA to total Dicer mRNA in cancer is higher than in adjacent normal mucosa, (iii) miR-103 in cancer is expressed similarly to adjacent normal mucosa, (iv) neither total Dicer mRNA expression nor long 3′UTR Dicer mRNA in cancer is correlated with miR-103 expression in cancer and clinicopathological features, such as TNM stage, tumor location, histological grade, and vessel invasion. Expression of Dicer varies among tumor types and may be correlated with cancer progression. Upregulation of Dicer was found in prostate, esophageal, and oral cancer,26, 27, 28 whereas Dicer downregulation was found in neuroblastoma, breast, ovarian, and advanced lung cancer.5, 29, 30, 31 Opinions are divergent on Dicer mRNA expression in CRC.13, 32, 33 Ciosea et al. reported that Dicer mRNA expression decreased in CRC,32 whereas Papachristou et al. found that Dicer mRNA levels did not display any significant differences between normal epithelium and CRC.33 Furthermore, Stratmann et al. found a significant increase of Dicer mRNA expression in the primary tumor from rectal cancer patients compared with normal mucosa but not in CRC patients.13 We have demonstrated that Dicer mRNA is downregulated in cancer compared with that in normal mucosa in agreement with the work by Ciosea et al. Therefore, it is possible that dysregulation of Dicer mRNA expression is induced by genomic instability at the Dicer loci at 14q, which are frequently observed to have loss of heterozygosity in CRC.34, 35 Further studies are needed to clarify the genetic abnormality at the Dicer loci. The Dicer mRNA has two polyadenylation sites, but the regulation of which site is used remains unknown. Recent researches suggest that polyadenylation site choice can be influenced by physiological conditions including cell growth, differentiation, developmental stage, and pathological events such as cancer.21, 22, 36 Mayr and Bartel found that cancer cell produce more mRNAs with short 3′UTR, including for Dicer, which indicates that the proximal polyadenylation site is used more frequently.15 The results of the current study are contrary to those of Mayr and Bartel and one reason may be the low transcriptional activity of Dicer in CRC. In a recent investigation into the 3′end-processing of mRNAs, short 3′UTR isoforms were found to be relatively more abundant when genes were highly expressed but conversely long 3′UTR isoforms were more abundant when genes were expressed at low levels.37 Elongated transcripts have also been observed in a mouse model of B-cell leukemia/lymphoma.22 Our findings are consistent with this observation that expression of long 3′UTR Dicer mRNA is higher relative to short 3′UTR Dicer mRNA in cancer than in normal mucosa and is associated with the downregulation of total Dicer mRNA expression in cancer. The expression of many miRNAs, including miR103/107, is affected to a lesser extent by Dicer expression because it only regulates a subset of miRNAs.16, 38, 39 As a result, the high level of miR-103/107 expression in breast cancer induces a reduction in the level of Dicer mRNA and results in the global downregulation of all miRNAs. There is a significant association between miR-103/107 expression and Dicer protein expression, and clinical relapse.16 miR-103/107 binds to the 3′UTR of Dicer mRNA, and inhibits protein expression by a posttranscriptional mechanism. Interestingly, the long 3′UTR Dicer mRNA has miR-103/107-binding sites in its 3′UTR but these are not present in the short 3′UTR Dicer mRNA. Therefore, the short 3′UTR Dicer mRNA is not directly regulated by miR-103/107. In the present study, miR-103 levels in CRC were not significantly different from normal mucosa. miR-103 levels in cancer cells might be sufficient to produce a reduction in the long 3′UTR Dicer mRNA, which results in the posttranscriptional downregulation of total Dicer mRNA, however, there might be other mechanisms to make a reduction in the long 3′UTR Dicer mRNA. Contrary to previous observations, Dicer mRNA expression was not associated with stage, tumor location, and histological grade in the current study. In terms of stage, the downregulation of Dicer mRNA might be an early event in colon carcinogenesis. In addition, long 3′UTR Dicer mRNA expression was not correlated with the same clinicopathological features. It is possible that this discrepancy is a result of the small population size of the study, which could be addressed with a larger number of cases in subsequent studies. We can conclude that the dysregulation of Dicer expression is involved in colon carcinogenesis through transcriptional and posttranscriptional gene regulation in CRC. Guarantor of the article: Shigeru Kanaoka, MD, PhD. Specific author contributions: Performed analysis: Yasushi Hamaya; performed research: Yasushi Hamaya; recruited patients to the study: Yasushi Hamaya, Shigeru Kuriyama, Tetsunari Takai, Ken-ichi Yoshida, Takanori Yamada, and Mitsushige Sugimoto; critically reviewed manuscript: Satoshi Osawa, Ken Sugimoto, Hiroaki Miyajima, and Shigeru Kanaoka. Financial support: This work was supported partly by a grant-in-aid from the Japanese Ministry of Education, Culture, and Science (23590935). Potential competing interests: Shigeru Kanaoka and Yasushi Hamaya are entitled to a share of royalties received by the University on licensing intellectual property to Olympus. The terms of these arrangements are being managed by the University in accordance with its conflict of interest policies. We thank Dr Naoyuki Miura and Dr Satoru Kono for their valuable advice as well as Junko Hasegawa, Ayako Honma, and Airi Nakamura for their technical assistance.