Since they were discovered nearly 30 years ago, microRNAs have emerged as novel targets for cancer diagnostics and therapeutics.
Alessandro Lagana, PhD
Since they were discovered nearly 30 years ago, microRNAs (miRNAs) have emerged as novel targets for cancer diagnostics and therapeutics. These small, endogenous, noncoding functional RNAs were notably described in studies of a mutant Caenorhabditis elegans worm in 1993.1 The initial report of miRNA dysregulation contributing to B-cell chronic lymphocytic leukemia was published in 2002.2 Subsequent research findings have shown that miRNAs control the expression of target messenger RNAs (mRNAs) to perform such functions as regulating tumor cell growth, invasion, angiogenesis, and adaptation to the tumor microenvironment.3,4
MiRNAs are approximately 22 nucleotides long and primarily function by binding to complementary sequences in the 3’-untranslated region (UTR) of mRNA to regulate protein translation.5 Because the miRNA—mRNA binding site is only 6 to 8 base pairs long, most miRNAs target multiple different mRNAs, and each mRNA can have target sites for multiple miRNAs, creating redundancy in the molecular networks for controlling gene expression.6
Tumor types have distinct miRNA signatures that differentiate them from normal tissues and other types of cancer. MiRNA expression profiles reflect the developmental lineage and differentiation state of tumors.7 Many genetic alterations in cancer cells are miRNA mechanism specific (eg, alterations in target binding, processing, and posttranscriptional editing), and cancer cells commonly have binding site variation in the 3’UTR of the target mRNA.8 Investigators have identified several mechanisms underlying the abnormal expression of miRNAs, including epigenetic modification of miRNAs, deregulation of transcription factors that target miRNAs, and deletion or amplification of miRNA genes. (Figure).9
The dysregulation of miRNA in cancer has increased interest in identifying miRNA signatures to predict prognosis, evaluate drug efficacy, identify embryonic/developmental origin, and resensitize drug-resistant tumors. Several preclinical studies have also investigated methods of targeting miRNA dysregulation, although more research is needed to optimize outcomes in the clinical setting.
Prediction of Prognosis
Investigators have researched miRNAs as prognostic markers because of their high stability in clinical samples, their robust patterns of expression, and their presence in blood, urine, and other body fluids. A meta-analysis of 43 studies in 20 cancer types showed that increased expression of miR-21 and decreased expression of let-7 were the most common miRNAs associated with patient outcome.10 A recent study using quantitative real-time polymerase chain reaction (PCR) techniques confirmed that expression of miR-21 and let-7 were increased and decreased, respectively, in patients with breast cancer compared with control individuals and patients with benign breast lesions,11 and results from another study showed that the single-nucleotide polymorphisms rs1042713 in the ADRB2 gene and rs11292 in the 3’UTR of the HIF1AN gene, 2 genes targeted by let-7, were associated with breast cancer susceptibility and that rs1017105 in the CLDN12 gene was a significant predictor of disease-free survival in breast cancer.12 In another study, 17 of 35 highly expressed miRNAs associated with transforming growth factor β signaling were combined to generate a risk score that significantly predicted survival in patients with advanced non—small cell lung cancer.13
MiRNA signatures may also be able to identify lymph node metastasis, a strong predictor of survival outcomes in patients undergoing curative resection for pancreatic ductal adenocarcinoma (PDAC). Investigators identified a 6-miRNA risk prediction model that distinguished patients with lymph node metastasis in a training cohort (area under the curve [AUC], 0.84; 95% CI, 0.76-0.89), which was confirmed in an independent clinical validation cohort (AUC, 0.73; 95% CI, 0.64-0.82).14 Because advanced radiographic imaging technologies may not identify occult distant metastases in patients with resectable PDAC, identification of a miRNA biomarker signature could be useful in improving selection of surgical candidates.15 A study with this goal identified an optimized 4-miRNA signature with robust diagnostic accuracy for stage IV (ie, metastatic) PDAC in the training cohort (AUC, 0.84; 95% CI, 0.76-0.90) and the independent validation cohort (AUC, 0.82; 95% CI, 0.72-0.90).15
Several miRNA dysregulation-causing alterations in cancer may predict or measure response to therapies. The G allele of the LCS6 polymorphism in the KRAS let-7 binding site was associated with lack of response to EGFR-targeted therapy in patients with KRAS and BRAF wild-type metastatic colorectal cancer,16 and variations in the miRNA binding sites in the 3’UTR of base excision repair genes were associated with colorectal cancer prognosis and response to 5-fluorouracil—based chemotherapy.17
Similarly, investigators are studying whether the addition of RNA sequencing to a DNA sequencing platform can improve identification of effective treatments for patients with cancer. Investigators at the Icahn School of Medicine at Mount Sinai in New York, New York, developed a platform to identify targets for FDA-approved cancer drugs. A pilot trial in 64 patients with multiple myeloma showed that the majority were able to receive a drug that was not specifically approved for multiple myeloma based on the RNA profile of their cancer and that many benefited from this approach.18
“We expect RNA sequencing will play a larger role in the precise delivery of targeted drugs in oncology,” Alessandro Lagana, PhD, an assistant professor at Icahn, said.18
Embryonic/Developmental Origin of Tumor Type
MiRNA expression signatures often mirror the embryonic or developmental origin of the tumor type, which suggests their potential role in classification of tumors. Results from a study showed that miRNA expression had >90% accuracy in the classification of 22 different tumor types according to tissue origin.19 Luminal, basal-like, and HER2 breast cancers were shown to have distinct miRNA-target regulation patterns,20 and miRNA expression correlated with abnormalities in acute myeloid leukemia, such as t(11q23), isolated trisomy 8, and FLT3-ITD mutations.21
Furthermore, isolation of stem/progenitor cells from prostate tumors showed distinct miRNA patterns among CD44-positive, CD133-positive, integrin α2β1—positive, and side population cells, indicating that the patterns of miRNA expression reflect cellular populations within a tumor.22
MiRNAs may also predict metastatic events associated with migration of specific cell types, as endogenous miR-126, which is silenced in several types of cancer, was shown to regulate endothelial recruitment to metastases in breast cancer cells in vitro and in vivo,23 and lung tumor cells had lower expression of miR-223-3p and higher expression of miR-29a-3p than did tumor microenvironment cells in a study examining the cellular origins of these circulating cell-free miRNAs.24
Targeting Drug Resistance
Recent studies have suggested that miRNAs play a key role in the development of intrinsic and acquired drug resistance, and manipulation of miRNA expression may be an option for resensitizing tumors to therapy. Downregulation of miR-451 in peripheral leukocytes was demonstrated in patients with imatinib (Gleevec)—resistant chronic myeloid leukemia (CML) and corresponded with increased levels of MYC, which was shown to have a binding site in the promoter region of miR-451. The authors suggested that this regulatory loop formed by MYC and miR-451 could be a target for imatinib-resistant CML.25
Plasma exosomal miRNAs have been investigated in the prediction of castration resistance development in prostate cancer. Next-generation sequencing of plasma exosomal miRNAs followed by quantitative real-time PCR to validate candidate miRNAs identified 6 miRNAs that were differentially expressed between treatment-naïve prostate cancer and castration-resistant prostate cancer (CRPC), and 5 of these were different between CRPC and treated non-CRPC. Application of receiver operating characteristic curves showed that miR-423-3p had an AUC of 0.784 for prediction of CRPC from treatment-naïve prostate cancer and an AUC of 0.883 for prediction of CRPC from treated non-CRPC. The authors concluded that plasma exosomal miRNAs could serve as a biomarker to predict the occurrence of CRPC.26
However, challenges remain in the identification and use of miRNAs as noninvasive biomarkers. Selection of appropriate endogenous controls for miRNA quantification is difficult because circulating RNases eliminate many mRNA and rRNA species in the blood.6 Several factors, such as diet and treatment, can also affect levels of circulating miRNAs, and changes in miRNA expression in the blood are rapid and can be affected by something as simple as a traumatic venipuncture.6 The plasma levels of certain circulating miRNAs originating from myeloid and lymphoid cells have also been shown to vary with changes in blood cell counts.27
Because of the close relationship between abnormal miRNA expression and tumorigenesis, prognosis, drug response and resistance, and survival, miRNA-based gene therapy has emerged as a target strategy for cancer treatment. Strategies for miRNA-based treatment primarily fall into one of 3 categories: (1) import of exogenous tumor-suppressor miRNAs (eg, with mimics or viral transfection) to inhibit proliferation or induce apoptosis of tumor cells; (2) inhibition of oncogenic miRNA function with antisense oligonucleotides strategies such as miRNA sponges, or miRNAmasking antisense oligonucleotides); or (3) use of artificial miRNAs designed to target 1 or more genes related to malignant tumor phenotypes (Table).28
Although multiple preclinical studies have been performed on various miRNAs, difficulties with systemic delivery of miRNA-based therapies have limited their clinical success to date. Therefore, improvements in delivery techniques, such as lipid nanoparticles, and novel approaches for modifying the tumor microenvironment are current focuses of research to optimize the efficacy of miRNA-based treatment for cancer.
Tumor Suppressor miRNAs
Tumor suppressor miRNAs are generally downregulated in cancer cells; therefore, investigators often aim to add specific miRNAs to reduce the activity of pathways involved in tumor growth, angiogenesis, and metastasis. Overexpression of the miR-892b decreased tumor growth, metastatic capacity, and angiogenesis through attenuation of NF-κB signaling in human breast cancer specimens,29 and re-expression of miR-625 suppressed cell migration and invasiveness through regulation of the IGF2BP1/PTEN pathway in hepatocellular carcinoma specimens.30 Results of a phase I study also showed that MRX34, a liposomal mimic of miR-34a (which downregulates the expression of >30 oncogenes) had acceptable safety and antitumor activity in a subset of patients with various types of solid tumors, including hepatocellular carcinoma.31
Modulation of the tumor microenvironment with tumor suppressor miRNAs has also been investigated recently. INT-1B3 (InteRNA) is a lipid nanoparticle-formulated, chemically modified miRNA 193a-3p mimic that inhibited tumor regrowth, reduced development of lung and intraperitoneal metastases, and induced longterm immunity in newly injected 4T1 tumor cells in a syngeneic orthotopic mouse model of triple-negative breast cancer (TNBC 4T1).32 Depletion of CD8-positive T cells significantly reduced protection against rechallenge with 4T1 tumor cells, suggesting long-term immune protection that was CD8 positive, T cell dependent. In addition, transfection of INT-1B3 inhibited tumor cell proliferation and survival, modulated the tumor cell cycle, induced apoptosis and/or senescence, and decreased tumor cell migration and invasion in a wide panel of human tumor cell lines. Furthermore, INT-1B3 showed significant antitumor activity in multiple human xenografts and an extended panel of syngeneic subcutaneous mouse tumor models.32
“We are convinced that INT-1B3, next to its potential in combination regimens, represents a unique and novel monotherapeutic opportunity in immuno-oncology and we look forward to evaluating its efficacy in [patients with] cancer,” said Roel Schaapveld, PhD, MBA, CEO of InteRNA, in a press release.32
Use of RNA-based therapy aims to address some of the shortcomings of current immunotherapeutic approaches, such as the inability of checkpoint inhibitors to target multiple immune checkpoints, the off-target adverse effects of their systemic administration, and the inability to inhibit immunosuppressive signals in the tumor microenvironment with chimeric antigen receptor T-cell therapy. To address these issues, investigators at RXi Pharmaceuticals created self-delivering RNA interference (sd-rxRNA) that silences multiple intra- and extracellular immunosuppressive targets by destroying RNAs before they are translated into the proteins that comprise the immune checkpoints.33 The investigators are currently exploring adoptive cell transfer, which involves ex vivo processing of immune cells and the use of sd-rxRNAi compounds to eliminate expression of immunosuppressive receptors or proteins from the therapeutic immune cells, thereby reducing their sensitivity to mechanisms of tumor resistance and improving their ability to destroy tumor cells. According to investigators, treating immune cells ex vivo before reinfusing them back to the patient has the potential to reduce immune-related adverse effects and target a wider range of immune checkpoints than conventional systemic immune checkpoint therapy.33
Because their overexpression has been demonstrated in a wide range of cancers, oncogenic miRNAs such as miR-21 and miR-155 have been identified as potential therapeutic targets for miRNA-based therapy. In vivo delivery of miR-21 inhibitors led to antitumor activity in immunodeficient mice harboring human xenografts of multiple myeloma,34 and transfection of an antisense oligonucleotide against miR-155 (typically oncogenic in breast cancer) into MDA-MB-157 breast cancer cells inhibited cell proliferation and increased cell apoptosis.35
However, many studies focus on a single miRNA, which may provide limited efficacy for treatment of cancer. One group of investigators synthesized artificially designed interfering long noncoding RNAs (i-lncRNAs) by generating tandem sequences of 10 copies of the complementary binding sequences of 13 oncogenic miRNAs. These i-lncRNAs compete with the corresponding mRNAs of oncogenic miRNA target genes, thereby “consuming” the oncogenic miRNAs and protecting tumor suppressor genes. Results from in vitro experiments showed that i-lncRNA expression inhibited cell proliferation and induced cell cycle arrest and apoptosis in diffuse large B-cell lymphoma (DLBCL) cell lines in addition to inhibiting the growth of DLBCL xenografts in nude mice.36 The authors concluded that targeting the carcinogenic effects of multiple oncogenic miRNAs that are not present in normal cells could improve antitumor efficacy and safety.36
Artificial miRNAs (amiRNAs) use natural miRNA precursor structures and interfere with expression of target genes by replacing core sequences of pre-miRNA with complementary sequences of target genes. These amiRNAs are regulated by the polymerase II promoter to achieve tissue-specific expression, potentially resulting in a favorable safety profile.28 Results from an in vitro study showed that introduction of the amiRNA p-27-5p into breast cancer cells led to inhibition of cell proliferation and cell cycle arrest via downregulation of CDK4 expression and consequent suppression of retinoblastoma protein (RB1) phosphorylation.37 In another study, miR-155-based amiRNAs against heparanase, an enzyme involved in tumor cell invasion and migration, reduced the adhesion, migration, and invasion ability of heparanase amiRNA-transfected melanoma cells and inhibited the mRNA and protein expression of IL-8 and chemokine (C-X-C motif) ligand 1 (CXCL1).38 However, the investigators acknowledged that improved delivery systems, such as viral vectors and modified liposomes, are needed to improve the delivery of amiRNAs in vivo.38
Although preclinical research on miRNAbased therapeutics is promising, suboptimal delivery methods and bioavailability have currently limited the success of these therapies in the clinical setting. Therefore, future research to improve delivery and target multiple miRNAs will likely improve the efficacy and safety of miRNA-based therapies.