MicroRNAs (miRNAs) are ∼22 nucleotide non-coding RNA molecules that regulate gene expression post-transcriptionally. Although aberrant expression of miRNAs in various human cancers suggests a role for miRNAs in tumorigenesis, it remains largely unclear as to whether knockdown of a specific miRNA affects tumor growth. In this study, we profiled miRNA expression in matched normal breast tissue and breast tumor tissues by TaqMan real-time polymerase chain reaction miRNA array methods. Consistent with previous findings, we found that miR-21 was highly overexpressed in breast tumors compared to the matched normal breast tissues among 157 human miRNAs analysed. To better evaluate the role of miR-21 in tumorigenesis, we transfected breast cancer MCF-7 cells with anti-miR-21 oligonucleotides and found that anti-miR-21 suppressed both cell growth in vitro and tumor growth in the xenograft mouse model. Furthermore, this anti-miR-21-mediated cell growth inhibition was associated with increased apoptosis and decreased cell proliferation, which could be in part owing to downregulation of the antiapoptotic Bcl-2 in anti-miR-21-treated tumor cells. Together, these results suggest that miR-21 functions as an oncogene and modulates tumorigenesis through regulation of genes such as bcl-2 and thus, it may serve as a novel therapeutic target.
MicroRNAs (miRNAs) are a class of naturally occurring small non-coding RNAs that control gene expression by targeting mRNAs for translational repression or cleavage (Pillai, 2005; Zamore and Haley, 2005). It is predicted that miRNAs comprise 1–5% of animal genes (Berezikov et al., 2005). miRNAs are transcribed as long primary transcripts in the nucleus and are subsequently cleaved to produce stem loop structured precursor molecules of ∼70 nt in length (pre-miRNAs) by Drosha (Kim, 2005), which are then exported to the cytoplasm, where the RNase III enzyme Dicer further processes them into mature miRNAs (∼22 nucleotides). Thus, miRNAs are related to short interfering RNAs, but they have distinct pathways (Bartel, 2004; Fitzgerald, 2005). Since the discovery of lin-4 in Caenorhabditis elegans (Lee et al., 1993; Wightman et al., 1993), thousands of miRNAs have been identified to date in a variety of organisms (http://microrna.sanger.ac.uk/).
As a new layer of gene regulation mechanism, miRNAs have diverse functions, including the regulation of cellular differentiation, proliferation and apoptosis (Chen et al., 2004; Croce and Calin, 2005). Thus, deregulation of miRNAs would alter the normal cell growth and development, leading to a variety of disorders including human cancer. For instance, about 65% of investigated patients suffering from B-cell chronic lymphocytic leukemia (CLL) have been reported to show a deletion located at chromosome 13q14 where the miR-15 and miR-16 genes are located and are under-represented in many B-CLL patients (Calin et al., 2002). Of interest, miRNA-containing regions are often located at fragile sites or in repetitive genomic sequences (Calin et al., 2004). Deregulation of other miRNAs has also been reported in different cancers (Michael et al., 2003; Metzler et al., 2004; Eis et al., 2005), indicating that there is a direct correlation between aberrant expression of miRNAs and human malignancy. However, although miRNAs have been the object of extensive research in recent years, the molecular basis of miRNA-mediated gene regulation is not fully understood and their role in tumorigenesis remains largely to be determined yet.
In this study, we found that miR-21 was overexpressed in breast tumor specimens, consistent with the previous report (Iorio et al., 2005). Importantly, anti-miR-21 oligonucleotides suppress both cell growth in vitro and tumor growth in vivo, which is associated with increased apoptosis and downregulation of the antiapoptotic protein Bcl-2.
Results and discussion
miR-21 is overexpressed in breast tumor tissues compared to matched normal breast tissues
Previous studies have shown that several miRNAs are aberrantly expressed in various types of cancers by miRNA array or Northern blot (Calin et al., 2002, 2004; Michael et al., 2003; Metzler et al., 2004; Eis et al., 2005). In this study, we profiled miRNA expression in matched normal breast and breast tumor tissues by TaqMan real-time polymerase chain reaction (PCR) using a newly released miRNA array from ABI (Forest City, CA, USA). The array carries specific primer sets that allow for detection of 157 mature human miRNAs. This method uses stem–loop reverse transcription (RT) primers; it is specific for detection of mature miRNAs (Chen et al., 2005; Lao et al., 2006). Furthermore, this method is very sensitive and is able to analyse miRNA expression in a single cell (Tang et al., 2006). We used U6 RNA for normalization of expression in different samples. From a total of five pairs of matched advanced breast tumor tissue specimens, miR-21 was the most abundantly expressed miRNA among all miRNAs in this array and moreover, the level of miR-21 was much higher in the tumor tissues than in the matched normal tissues (Figure 1a). As one CT (threshold cycle) unit is equivalent to ∼2-fold difference (Chen et al., 2005), this conversion would result in over a five-fold increases in miR-21 levels for tumor tissues compared to the matched normal tissues after normalization to U6 RNA (Figure 1b), consistent with the previous report (Iorio et al., 2005). Furthermore, using the individual miR-21 primer set, we were able to confirm these results in more matched breast tumor samples (not shown).
Anti-miR-21 inhibits cell growth in vitro
To test whether miR-21 may function as an oncogene, we examined the effect of suppression of miR-21 on breast tumor cell growth. Thus, we used anti-miR-21 inhibitor as this approach has been successfully used to inhibit miR-21 (Chan et al., 2005; Cheng et al., 2005). The anti-miR-21 inhibitor is a sequence-specific and chemically modified oligonucleotide to specifically target and knockdown miR-21 molecule. TaqMan real-time PCR revealed that anti-miR-21 significantly reduced miR-21 level (Figure 2a), suggesting that anti-miR-21 is efficiently introduced into the cells and knock down miR-21. This is probably due to the formation of highly stable complexes of miR-21 with anti-miR-21 that prevents miRNA detection by TaqMan real-time PCR. Of interest, we found that anti-miR-21 reduced cell growth in a dose-dependent manner. At 50 nM, the growth inhibition by anti-miR-21 reached about 25%, at day 3 after transfection (Figure 2b); this result was also in agreement with the previous report that miR-21 inhibitors decrease human glioblastoma cell survival (Chan et al., 2005). To further assess the effect of anti-miR-21 on cell growth, we treated the transfected cells with the anticancer drug topotecan (TPT) that is known to inhibit DNA topoisomerase I and cause DNA damage (Tanizawa et al., 1994). Anti-miR-21-mediated cell growth inhibition was increased up to 40% when the transfected cells were treated with 0.1 μM TPT (Figure 2c). Therefore, anti-miR-21 can inhibit cell growth in vitro. These results also suggest that suppression of miR-21 can sensitize tumor cells to anticancer agents.
Anti-miR-21 inhibits tumor growth in the xenograft carcinoma mouse model
Although it has previously been shown that there is a direct correlation between aberrant expression of miR-21 and breast cancer (Iorio et al., 2005), it is not clear whether suppression of miR-21 alone will affect tumorigenesis. Therefore, we transiently transfected MCF-7 cells with anti-miR-21 or the negative control, and then injected them into mammary pads of female nude mice. Of considerable interest, we found that tumors derived from MCF-7 cells transfected with anti-miR-21 grew substantially slowly, compared to the negative control during the whole tumor growth period (Figure 3a). By day 28 when tumors were harvested, average weight for tumors derived from cells transfected with anti-miR-21 was only about half of those derived from the cells transfected with the negative control (Figure 3b). Immunostaining with the anti-Ki-67 indicated that the reduced tumor growth is likely due to a lower proliferation caused by anti-miR-21 because Ki-67 staining was much weaker for anti-miR-21 than for the negative control (Figure 3c). These results strongly suggest that miR-21 plays an important role in tumorigenesis. To test how long suppression of miR-21 by anti-miR-21 in tumors can sustain, we measured the miR-21 levels. We found that the suppression effect lasted up to 2 weeks (Supplementary materials), suggesting that the initial suppression of miR-21 is sufficient to inhibit tumor growth.
Of interest, the inhibitory effect of anti-miR-21 on tumor growth (Figure 3b) is greater (∼50%), compared to its inhibitory effect on cell growth in vitro (∼25%) (Figure 2b). Although the observation time for tumor growth is longer than in vitro cell growth inhibition assays, which could explain in part the difference, other factors could also contribute to this difference. For instance, stress from the tumor microenvironment, such as hypoxia, may enhance the inhibitory effect of the anti-miR-21. This appears to be in agreement with the finding that other stresses, such as DNA damage caused by TPT, can increase the inhibitory effect mediated by anti-miR-21 (Figure 2c). Alternatively, anti-miR-21 could also affect genes that are linked to other tumorigenesis factors, which might explain in part why more inhibition for anti-miR-21 was seen in tumors than cell growth in vitro.
Anti-miR-21 increases cell apoptosis which is associated with downregulation of bcl-2 expression
To dissect the molecular basis underlying this miR-21-associated alteration of tumor growth, we searched for potential miR-21 targets using programs available (e.g., http://microrna.sanger.ac.uk/targets/v2/; http://genes.mit.edu/cgi-bin/targetscan_lookup2.pl?KEYWORD=miR-21) and tested several genes that are likely involved in tumorigenesis, such as FasL. However, they were not affected by anti-miR-21 (Supplementary materials). Thus, we tested whether anti-miR-21 suppresses cell growth by triggering apoptosis pathways as previous studies have suggested that miR-21 regulates apoptosis pathways in tumor cells (Chan et al., 2005). Consistent with the previous report for glioblastoma cells (Chan et al., 2005), but contrary to the results in HeLa cells (Cheng et al., 2005), we found that anti-miR-21 caused more apoptosis than the negative control in MCF-7 cells by a 4.5-fold (Figure 4a). To further determine the possible involvement of apoptosis in anti-miR-21-mediated growth inhibition, we treated transfected cells with the general caspase inhibitor Z-VAD-fmk. As shown in Figure 4b, Z-VAD-fmk was able to reverse the growth inhibition caused by anti-miR-21, suggesting that increased apoptosis in the anti-miR-21-treated MCF-7 cells is at least in part responsible for the observed growth inhibition. Furthermore, we detected a lower level of Bcl-2 protein in the anti-miR-21-transfected MCF-7 cells (Figure 4c and d) as well as tumors derived from the MCF-7 cells transfected with anti-miR-21 (Figure 4e). Given that suppression level of Bcl-2 in vivo (Figure 4e) is greater than that in vitro (Figure 4c), it is possible that tumor microenvironment may enhance downregulation of Bcl-2 in the anti-miR-21-treated tumors. We also tested other apoptosis-related proteins such as p53 and PUMA, and found no difference between the negative control and anti-miR-21 (Supplementary materials). Thus, the induction of apoptosis by anti-miR-21 is possibly in part owing to downregulation of Bcl-2. We also examined bcl-2 mRNA by RT–PCR and found that bcl-2 mRNA was decreased in the anti-miR-21-treated cells (Supplementary materials), suggesting that miR-21 may regulate bcl-2 expression indirectly. Although we cannot exclude the possibility that anti-miR-21 may cause degradation of bcl-2 mRNA, one possibility would be that anti-miR-21 suppresses expression of a gene(s) that negatively regulates bcl-2 expression. Therefore, identification of direct miR-21 targets may provide new insight into how miR-21 controls expression of genes involved in apoptosis pathways including bcl-2.
In summary, we show that miR-21 is overexpressed in breast tumor tissues and anti-miR-21 inhibits both cell growth in vitro and tumor growth in vivo. This is possibly owing to increased apoptosis associated with downregulation of bcl-2 expression. As experiments with the xenograft carcinoma model indicate that one transient transfection with anti-miR-21 is sufficient to cause substantial inhibition of tumor growth, this raises the possibility that anti-miR-21 may have potential therapeutic value. Indeed, anti-miRNA oligonucleotides can stay a relatively long period of time in animals (Krutzfeldt et al., 2005). Therefore, miRNAs, in particular miR-21, may serve as potential targets for cancer therapy.
We are grateful to CHTN and SIU tumor bank for providing patient specimens. We thank Rupinder Grewal and Heather Mizeur for cutting frozen tumor samples. This study was supported in part by Grants CA102630 from NCI and BC045418 from DOD.