The clinical potential of microRNAs

Multiple reports have noted the utility of microRNAs for the diagnosis of cancer [37, 38]. microRNA expression profiles have been used to distinguish tumor from normal samples, identification of tissue of origin for tumors of unknown origin or in poorly differentiated tumors and to distinguish different subtypes of tumors. Sample datasets have been stratified to show that certain alterations of microRNAs occur in patients at an early stage of cancer and thus may be quite useful for early detection. Large tissue specimens are not needed for accurate MicroRNA detection since their expression can be easily measured in biopsy specimens. Although the majority of these studies have used tissue to assess microRNA levels, recent studies have shown that microRNAs can be measured in formalin fixed paraffin embedded (FFPE) tissues [39]. Given the invasive nature of fresh/frozen tissue collection and the availability of FFPE, this serves as a major advance in the feasibility of measuring microRNA levels for the purposes of diagnosis. Recent studies have also shown that microRNAs can be detected in serum. These studies offer the promise of utilizing microRNA screening via less invasive blood-based mechanisms. In addition, mature microRNAs are relatively stable. These phenomena make microRNAs superior molecular markers and targets for interrogation and as such, microRNA expression profiling can be utilized as a tool for cancer diagnosis [10, 40].

The potential clinical utility of microRNA extends beyond the realm of diagnosis to other important clinical measures such as prognosis and treatment response. A series of publications has shown that microRNAs are useful indicators of clinical outcome in a number of cancer types [10, 40‐45]. In addition, microRNAs have been shown to play a predictive role in determining the tendency for recurrence and metastasis [46‐50]. These microRNA alterations have not only been found in tumor specimens, but have also been observed in surrounding non-cancerous tissue, indicating that microRNAs may also serve to detect alterations in the cancer microenvironment [45, 51, 52]. microRNAs have also been shown as useful indicators of which patient groups may respond better to a particular treatment regimen. An example of this was shown for liver cancer patients, whereby miR-26 expression could be used to stratify patients for IFN treatment [53]. The full potential of microRNAs as prognostic factors awaits the results of larger prospective studies.

The rules of Watson and Crick base-pairing guide the binding of microRNAs to their target sites. In order to circumvent this interaction, anti-microRNA oligonucleotides (AMOs) have been generated to directly compete with endogenous microRNAs [54]. However, the ability of AMOs to specifically inactivate endogenous targets has been shown to be quite inefficient. Thus, several modifications of AMOs have been generated to improve their effectiveness and stability such as the addition of 2'-O-methyl and 2'-O-methoxyethyl groups to the 5' end of the molecule [55]. Studies have shown that targeting of miR-21, a microRNA that is overexpressed in many cancer types, by such methods effectively reduced tumor size in a xenograft mouse model based on MCF-7 cells [56]. AMOs conjugated to cholesterol (antagomirs) have been also been generated and have been described to efficiently inhibit microRNA activity in-vivo [57]. In addition, locked-nucleic-acid antisense oligonucleotides (LNAs) have been designed to increase stability and have been shown to be highly aqueous and exhibit low toxicity in-vivo [58]. In gliomas, this method has been effectively used to completely eradicate miR-21 [59]. Another method for reducing the interaction between microRNAs and their targets is the use of microRNA sponges. These sponges are synthetic mRNAs that contain multiple binding sites for an endogenous microRNA. Sponges designed with multimeric seed sequences have been shown to effectively repress microRNA families sharing the same seed sequence [60]. Although microRNA sponges perform as well as chemically modified AMOs in-vitro, their efficacy in-vivo remains to be determined.

Although these oligonucleotide-based methods have been shown to work, they do elicit off-target side effects and unwanted toxicity. This is due to the capability of microRNAs to regulate hundreds of genes. A strategy called miR-masking is an alternative strategy designed to combat this effect. This method utilizes a sequence with perfect complementarity to the target gene such that duplexing will occur with higher affinity than that between the target gene and its endogenous microRNA. The caveat of this approach is that the choice of target gene must be specific in order to effectively reduce the interaction. This gene-specific, microRNA interfering strategy has been shown to reduce the activities of miR-1, miR-133 and miR-430 in several model systems [61, 62]. Another strategy to increase specificity of effects is the use of small molecule inhibitors against specific microRNAs. Azobenzene, for example, has been identified as a specific and efficient inhibitor of miR-21 [63]. Although the effectiveness of such inhibitors awaits exploration in-vivo, they are potentially promising tools for cancer therapy.

Elevating the expression of microRNAs with tumor suppressive roles is a strategy to restore tumor inhibitory functions in the cell. This can be achieved through the use of viral or liposomal delivery mechanisms [64, 65]. Several microRNAs have been introduced to cells via this methodology, including miR-34, miR-15, miR-16 and let-7 [66‐69]. Systemic administration of miR-26, a tumor suppressive microRNA in HCC, using adenovirus-associated virus (AAV) in an animal model of HCC, results in inhibition of cell proliferation and tumor-specific apoptosis [70]. This approach reduces toxicity since AAV vectors do not integrate into the host genome and eventually are eliminated. Although viral vector-directed methods show high gene transfer efficiency, they lack tumor targeting and residual viral elements can elicit immunonogenic effects. This has led to the development of non-viral methods of gene transfer such as cationic liposome mediated systems. These lipoplexes are promising, but they lack tumor specificity and have relatively low transfection efficiency when compared to viral vectors.

MicroRNA mimics have also been used to increase microRNA expression. These small, chemically modified double-stranded RNA molecules mimic endogenous mature microRNA. These mimics are now commercially available and promising results have been reported with systemic delivery methods using lipid and polymer-based nanoparticles [71‐73]. Since these mimics do not have vector-based toxicity, they are promising tools for therapeutic treatment of tumors.