miR-22 AS AN ONCOGENE
The tumor suppressor PTEN (phosphatase and tensin homolog deleted from chromosome 10), is a lipid phosphatase that converts phosphatidylinositol-3, 4, 5- triphosphate (PIP3) into phosphatidylinositol (4, 5)- diphosphate (PIP2). PTEN is well-known as the most highly mutated tumor suppressor gene in the p53-post era [66, 67]. Recently, three papers by Nadav Bar and Rivka Dikstein, Linhua Liu and colleagues, and Laura Poliseno and colleagues, have demonstrated that PTEN was a bona fide target of miR-22 in a small cohort of cancer cell lines that are driven from breast cancer, cervical cancer, prostate cancer, and bronchial epithelial cancer. These studies contradict a uniform role of miR-22, indicating that under certain circumstances, miR-22 may function as an oncogene because of its antagonistic effects on tumor suppressive PTEN signaling [26, 68, 69] (Fig. 3).
Using various miRNA target prediction programs and/or RNAhybrid program for evaluating the minimum free energy hybridization, these three groups all found that miR-22-PTEN was a high scoring miRNA-target pair. Enforced or reduced expression of miR-22 in human HEK293T, cervical cancer HeLa and breast cancer MCF-7 cell lines (Nadav Bar and Rivka Dikstein), anti-benzo[a]pyrene-7, 8-diol-9, 10-epoxide (anti-BPDE)-induced transformed human bronchial epithelial cancer cell line 16HBE-T (Linhua Liu and colleagues) and prostate cancer cell line DU145 (Laura Poliseno and colleagues), revealed that miR-22 negatively regulated PTEN protein expression. Intriguingly, an inverse correlation between miR-22 and PTEN mRNA expression has been presented by Poliseno et al., while Liu et al. found that there was no change of PTEN mRNA expression regardless of miR- 22 levels. A similarly inverse association of miR-22 and
J. Xiong
PTEN protein levels was observed in 16HBE-T and its parental normal cell line16HBE (Linhua Liu and colleagues), and in several prostate cancer cell lines, prostate cell lines and a prostate tumor tissue microarray (Laura Poliseno and colleagues). Furthermore, the mature levels of miR-22 were significantly increased in these tumor cells versus their normal counterparts. In line with this result, a direct correlation between miR-22 expression and phosphorylated AKT or between the expression of miR- 22 and that of DICER was also identified by Laura Poliseno and colleagues. These three groups all showed that an intact binding site at the 3’UTR of PTEN mRNA was required for miR-22 targeting. Functional analyses showed that miR-22 could induce apoptosis, inhibited colony formation and suppressed motility of bronchial epithelial cancer cells (Linhua Liu and colleagues), and intrinsically promote prostate cancer cell growth and tumorigenesis in tumor-bearing nude mice (Laura Poliseno and colleagues). Subsequently, the influence of miR-22 on the downstream signaling of PTEN was tested. Bar et al. showed that miR-22 could stimulate AKT activity, and in turn, AKT significantly upregulated miR-22 expression, suggesting a regulatory loop comprising miR-22, PTEN and AKT. Analogously, Poliseno et al. showed that miR-22-mediated oncogenic activity was dependent on decreased PTEN expression and increased phosphorylated AKT [26, 68, 69].
Surprisingly, the conclusion presented by these three papers is challenged by a more recent study that argues against the downregulation of PTEN by miR-22 [70]. The cytotoxicity of paclitaxel, a featured anti-tumor activity, can induce apoptosis, and inhibit proliferation and survival of p53 wild-type colon cancer cells, rather than chemoresistant p53-mutated colon cancer cells.
Chemoresistance assay showed that overexpression of miR-22 prevented the chemoresistance to paclitaxel and induced apoptosis, and inhibited proliferation and survival of p53-mutated colon cancer cells. Finally, Li et al. pinpointed that tumor-suppressive miR-22 decreased AKT activity and MTDH expression, and subsequently increased Bax and active caspase-3 expression through upregulation of PTEN expression. These studies emphasize the complex of miR-22 function in tumor pathways, and provide another good example identical to miR-17-92 cluster that functions as either tumor suppressors or oncogenes due to cell types and expression profiles of miRNA-target pairs [71]. Paradoxically, contrary to the conventional view that different target repertoire of one miRNA contributes to its distinct functional role, miR-22 leads to the opposite effects on the regulation of the same target gene PTEN and the acquisition of tumor cell phenotypes. Further investigations will be required to elucidate this cellular context-dependent role of miR-22 as a key negative or positive regulator of PTEN, and as a tumor suppressor or an oncogene.
3. miR-22 AS A TUMOR SUPPRESSOR
At least 9 genes including estrogen receptor alpha (ER), histone deacetylase 4 (HDAC4), CDK6, SIRT1, Sp1, p21, hypoxia inducible factor 1 (HIF-1), MYCBP and MAX, and 2 candidate genes (EVI1 and EZR) have been reported so far to be the direct targets of miR-22 for its anti-tumorigenic capacity in 10 types of cancer (leukemia, lymphoma, breast cancer, lung cancer, hepatoma, pancreatic cancer, ovarian cancer, colon cancer, cervical cancer, and prostate cancer). Many other potential targets of tumor-suppressive miR-22 have been revealed by microarray- or beads-based detection, combined with computational predictions but await further validation. Downregulation of these targets by miR-22 was accompanied by slow cell proliferation, smaller and fewer colony formation in soft agar, inhibited tumorigenesis, suppressed cell motility, apoptosis, senescence, anti-angiogenesis, and/or cell cycle arrest [9, 34-51] (Fig. 2).
The anti-tumor properties of miR-22 was first to be demonstrated in curcumin (diferuloylmethane)
e rhizome of Curcuma longa, is a promising chemotherapeutic drug candidate for its low toxicity and robust anti-tumor activity in preclinical trials. However, the mechanism by which curcumin inhibits tumor growth is still obscure. To address this issue, Sun et al. have developed a home-made microarray chip to profile miRNA expression in curcumin-treated pancreatic cancer cells, and reported that the deleterious effect of curcumin on pancreatic cancer cells can be ascribed, at least in part, to its upregulation of miR-22 that reduced the expression of ER and Sp1 protein in flow cytometry [34]. ER is a key member of nuclear receptor family as a crucial player in tumorigenesis and a traditional biomarker for clinical diagnosis and prognosis [52]. A high-throughput protein lysate microarray (LMA) screen has identified miR-22 as a potent inhibitor of ER expression. Cell cycle and cell proliferation analyses showed that miR- 22 induced G0/G1 phase arrest and exhibited repressed growth of breast cancer cells [35]. Moreover, miR-22 degraded ER mRNA by binding the evolutionarily conserved target site at the whole ~4.3 kb 3’ UTR of ER, and attenuated ER-dependent growth of breast cancer cells through downregulation of ER expression [36]. Our group also independently identified ER as a direct target of miR-22 in breast cancer cells. Furthermore, we have assessed the expression of miR-22 in 5 breast cancer cell lines (ER-positive MCF-7, T-47D, BT-474, and ER- negative MDA-MB-231, SK-BR-3) and 23 clinical tissue specimens (10 are ER positive, and 13 are ER- negative). We found that miR-22 showed decreased abundance in ER-positive breast cancer cell lines and clinical specimens, in contrast with ER-negative counterparts. Indeed, ER participated in the inhibition of breast cancer cell proliferation and anchorage- independent growth by miR-22 [37]. Zhang et al. have carried out real-time quantitative reverse transcription PCR (qRT-PCR) to examine miR-22 expression in 160 paired hepatocellular carcinoma (HCC) biopsies and found that miR-22 expression was downregulated in HCC. As such, low expression level of miR-22 serves as a vital signature for diagnosis and poor survival in HCC patients. Furthermore, the cancer-development- associated HDAC4 was a direct target of miR-22 for its anti-tumor growth action in vitro and in vivo, and there was a significant inverse association between the miR- 22 levels and HDAC4 expression in HCC tissues [38]. A miRNA microarray performed by Dan Xu and colleagues showed that miR-22 was highly expressed in senescent human fibroblasts, but had greatly reduced expression in various human tumor cells. Ultimately, miR-22 exemplified a novel senescence- associated miRNA that induces cellular senescence in human normal fibroblasts and cancer cells by directly targeting senescence-associated genes CDK6, SIRT1, and Sp1. Consistently, overproduction of miR-22 could induce senescence-like phenotype in cervical cancer cells, and the miR-22-mediated senescence in a mouse model can lead to inhibition of breast tumor growth and metastasis in vivo [39]. In a more recent
study, p21 was shown to be a direct target of miR-22 during p53-dependent apoptosis of tumor cells. Using a custom 454 miRNAs-carrying expression library and MTS assays, Tsuchiya et al. have identified miR-22 as a potent suppressor of colon cancer cell proliferation. Comparative genomic hybridization array results supporting this notion showed that miR-22 had highly frequent hemizygous deletions in 24 colon cancer patients. Similarly, the expression of miR-22 was downregulated in six colon cancer cell lines in comparison with normal colon epithelial cells. Furthermore, miR-22 induced apoptosis in p53-wild type colon cancer cells by directly targeting p21 3’ UTR. In addition, ChIP and qRT-PCR assays showed that miR-22 gene acted as a direct transcriptional target of p53, and was under the tight transcriptional control of p53. Altogether, these findings indicated that p53 induced apoptosis through downregulation of miR- 22, a strong inhibitor of p21 [40]. Recently, Yamakuchi et al. identified miR-22 as an anti-angiogenesis factor in colon cancer. It has been shown that miR-22 was lowly expressed in colon cancer cell lines and specimens. Furthermore, miR-22 inhibited HIF-1 translation by targeting its 3’UTR, inducing suppressed tumor angiogenesis, and inhibited endothelial cell growth and invasion through downregulation of HIF-1 and VEGF expression [9].
Through interaction with partners, c-Myc impinges on a sophisticated network of target genes to induce tumorigenesis as a crucial oncogenic transcription factor [53-55]. Recent studies show that c-Myc modulates many miRNAs that are implicated in tumorigenesis, embryonic stem cell differentiation and glutamine metabolism [56-63]; however, how a miRNA controls c-Myc activity is still enigmatic. Although a myriad of c-Myc-mediated miRNAs confer altered phenotypes, c-Myc appears to escape miRNA- dependent silencing due to its extremely short 3’-UTR. Using expression profile microarray to explore the candidate targets of miR-22, we found that a c-Myc binding protein-coding gene, MYCBP, was dramatically downregulated by miR-22. Furthermore, miR-22 suppresses breast cancer cell growth, at least in part, by targeting MYCBP 3’UTR and limiting MYCBP expression [41]. MYCBP is a small c-Myc binding protein of ~ 11kD that binds to the N-terminal domain of c-Myc via its C-terminal region and activates the transcription of the E-box-dependent c-Myc target genes for tumor growth [64, 65]. As expected, the expression of a panel of selected E-box-containing c- Myc target genes was diminished by miR-22 overexpression. These findings, considered together with c-Myc-mediated repression of miR-22 expression [59], led us to propose a novel positive feedback regulatory loop where repression of miR-22 by c-Myc enhanced MYCBP-mediated transcription of E-box- containing c-Myc target genes for tumorigenesis [41]. Supporting this hypothesis is the fact that Ting et al. have identified another c-Myc binding partner Max as a direct target of miR-22 in many forms of human cancer including leukemia, prostate cancer, lung cancer and
Insights into miR-22
Current Molecular Medicine, 2012, Vol. 12, No. 3 251
breast cancer. Cell cycle cessation was attributed to the upregulation of miR-22 by 12-O-tetradecanoyl- phorbol-13-acetate (TPA), a differentiation agent that can trigger the protein kinase c (PKC)-extracellular regulated kinase (ERK) signaling module for cell differentiation. Differentiation-associated miR-22 could suppress cancer cell growth and regulate the transcription of the downstream targets of Myc-Max complex [42]. Taken together, two c-Myc binding proteins Max and MYCBP have been characterized as direct targets of miR-22 in a large fraction of cancers. Therefore, it has been hypothesized that miRNAs exert their regulatory effects on c-Myc mainly in an indirect manner via perturbation of c-Myc binding proteins [41]. Moreover, based on independent databases and experimentally validated data, a list of mixed miRNA/transcription factor feed-forward loops orchestrated by Myc as master regulator has been constructed. This list provided first system models for the Myc/miRNAs cooperation in gene regulation at both transcriptional and post-transcriptional levels [43].
Consistent with above identified direct targets of miR-22, convincing evidence lends further support to the anti-tumor activity of miR-22. Using real-time qRT- PCR method, Xiaoqing Li and colleagues compared miR-22 transcription in acute lymphoblastic leukaemia (ALL) cell line NALM-6 and the peripheral blood mononuclear cells (PBMCs) from 7 healthy volunteers, 18 ALL patients and 9 acute myeloid leukaemia (AML) patients, and found that the histone deacetylase inhibitor, trichostatin A (TSA) elevated the expression of miR-22 gene specifically in NALM-6 and PBMCs from ALL patients compared with healthy volunteers, not in PBMCs from AML patients. Moreover, the combined bisulfite restriction analysis and the methylation-specific PCR assay in NALM-6 cells and PBMCs from ALL patients indicated no methylated CpG islands in the promoter element of miR-22 gene. These findings suggested that H3K27triM-associated histone modification at the promoter of element of miR- 22 gene in ALL cells led to DNA methylation- independent gene silencing of miR-22. This process allowed the functional failure of miR-22 as a potential tumor suppressor, and could be released by TSA [44]. Illumina next-generation sequencing technology with millions of small RNA beads was employed by Nagaraja et al. have identified approximately 7-fold decrease of miR-22 expression in human clear cell ovarian cancer cell lines compared with short-term primary cultures of human normal ovarian surface epithelium. Gene expression profiling showed that overexpression of miR-22 downregulated the expression of 22 predicted targets determined by TargetScanHuman (release 5.0; http://www.targetscan. org/), PicTar (http://pictar.mdc-berlin.de/) and miRanda (September 2008; http://www.microrna.org/). The repressed genes included an effector gene, EDC3, for P-body assembly and mRNA decapping during miRNA- mediating silencing, and multiple oncogenes, such as ARRB1, CLIP2, EVI1 and FRAT2. Enforced expression of miR-22 in clear cell ovarian cancer counteracted the
expression of tumor-promoting EVI1, which inhibited apoptosis and promotes cell survival by stimulating phosphatidylinositol 3-kinase (PI3K)/ v-akt murine thymoma viral oncogene (AKT) signaling. A dramatic shift of global gene expression pattern in human clear cell ovarian cancer could be provoked by miR-22 overexpression from cancerous status to a more normal status [45]. The inhibitory effects of miR-22 on oncogenic pathways and the propensity to reverse the malignant properties of cancerous tumors underlie the potential therapeutic benefits of miR-22 as an attractive anti-cancer agent. MiRNA expression profile showed that out of all measured 254 miRNA genes, miR-22 displayed the most significant repression of ~50% in metastatic clones (derived from bone, lung and adrenal) compared with parental MDA-MB-231 breast cancer cells (derived from primary tumor in the mammary fat pad of nude mice). Furthermore, the expression of ERBB3 and EVI1, and their downstream AKT signaling were repressed in the miR-22 overexpressing MDA-MB-231 cells. Remarkably, higher expression of miR-22 transcript correlated with better survival, and lower expression of EVI1 in primary tumors. These data have strengthened the view that metastatic breast cancer cells generated sufficient oncogenic EVI1 through downregulation of miR-22 [46]. The expression of miR-22 has been shown to be declined in highly metastatic ovarian cell lines. Crucially, cell migration and invasion both in gain-of- function and loss-of-function studies could be negatively regulated by miR-22. Further investigations revealed that miR-22 enabled the translational inhibition of metastasis-related gene EZR, and inhibited the expression of multiple predicted metastasis-related genes including ESR1, TIAM1, RHOC, RHOA, TP53, SNAI1 and SIX1 [47, 48]. Garzon et al. have developed a custom microarray platform to evaluate the miRNA expression pattern in 85 adult AML de novo patients bearing mutated NPM1 (55 cases) compared to the NPM1-unmutated cases (30 cases). Both this microarray analysis and qRT-PCR measurement identified that of all 21 downregulated miRNAs, miR-22 had the most differentiated expression of ~3 fold [49]. Intrahepatic cholangiocarcinoma (ICC) arises in the small bile ducts as the second most common primary liver cancer. Small RNA library sequencing and qRT- PCT detection have identified a unique signature of downregulated miR-22 expression in ICC cell lines compared to normal intrahepatic biliary epithelial cell line, HIBEpiC. These results suggested miR-22 as a potential biomarker for ICC diagnosis and prognosis [50]. Melo et al. reported that frameshift mutations of TARBP2 (TAR RNA-binding protein 2), which encodes an integral component of a DICER1-containing complex, induced TRBP impairment and decreased miR-22 expression in a colorectal cancer cell line Co115. These findings provided an explanation for defects of miR-22 in sporadic and hereditary carcinomas with microsatellite instability, and suggested the potential tumor-suppressing nature of miR-22 [51].
Paraphries these articles in your own words?
