Development of microRNA-based therapeutics for vascular disease

Posttranscriptional regulation of gene expression is an important mechanism by which diverse physiologic processes are modulated, including cell differentiation, growth, proliferation, and apoptosis. MicroRNAs (miRNAs) are a class of short, highly conserved, noncoding RNAs that posttranscriptionally regulate gene expression through interaction with the 3’ untranslated region (3’UTR) of target messenger RNA (mRNA).¹ These posttranscriptional regulators have become a major focus of molecular biology research over the past 15 years because of the pivotal roles miRNAs play in virtually all aspects of development and disease. MiRNAs recognize mRNA targets mainly by the miRNA seed sequence, typically nucleotides 2–8 of the 5’ end. However, an important feature of miRNA-mediated gene regulation is that mismatches in the 3’ sequence allow miRNAs to potentially target hundreds or even thousands of genes. This multitargeting feature opens up the possibility of one miRNA having the capability to regulate entire biological pathways, which, given that human disease often involves deregulation of multiple genes, has led to considerable interest in miRNAs as therapeutic targets. For example, in 2018, nearly 3,500 studies were published that had potential implications for miRNA-based therapies.² However, despite this considerable volume of studies, questions remain as to the actual clinical application of miRNA-based drugs and how far they truly are from the patients.

Context-dependent activity is another feature of miRNAs with important therapeutic implications; miRNA function can depend on cell type and/or cellular compartment. This was demonstrated over 10 years ago for one of the best-characterized oncogenic miRNAs, the miR-17-92 cluster, in regulation of angiogenesis. The miR-17-92 cluster is a prototypical polycistronic miRNA gene, encoding 6 mature miRNAs: miR-17, miR-18a, miR-19a, miR-20a, miR-19b, and miR-92a, all processed from a single primary transcript. In tumor cells, miR-17-92 was shown to have proangiogenic functions.³ However, Doebele and co-workers demonstrated that, within endothelial cells, individual members of the miR-17-92 cluster exhibited cell-autonomous, antiangiogenic activity.⁴ By showing that members of the cluster exerted distinct functions in physiologic versus pathologic models of angiogenesis, this work provided important insight into context-dependent differences in miRNA activity. From as therapeutic standpoint, this work highlighted the need for cell type-specific targeting strategies if the potential of miRNA-based therapeutics was going to be fully exploited.

In the current issue of Circulation Research, Marín-Ramos et al. explore the therapeutic potential of a member of the miR-17-92 cluster, miR-18a, in treating brain arteriovenous malformations (AVMs).⁵ An AVM is a vascular mass (i.e. nidus) where arteries are directly shunted to veins without an interposed capillary system. Brain AVM is a rare but important cause of intracerebral hemorrhage in young adults and is associated with high morbidity and mortality. Current treatments are highly invasive, not often feasible, and AVMs usually recur following treatment, so there is a need for novel non-invasive therapies. The current study extends previous work by this group that indicated therapeutic potential for miR-18a in treating brain AVMs; miR-18a treatment of brain endothelial cells (BECs) from AVMs improved the aberrant, proangiogenic phenotype demonstrated by these cells. In particular, miR-18a treatment, through decreasing inhibitor of DNA binding 1 (ID1) and increasing thrombospondin-1 (TSP-1), was shown to decreased VEGF (vascular endothelial growth factor)-A and VEGF-D secretion by these cells.⁶ In the present study, the authors delved deeper into the mechanistic underpinnings of miR-18a-mediated changes in AVM-BEC phenotype, as well as provide insight into miR-18a as a therapy for AVM. They show that miR-18a, delivered intranasally, can decrease abnormal cerebral vasculature in a mouse model of AVM.

Excessive VEGF activity is an important factor in the development and progression of AVMs. The authors identified the proangiogenic serpin, PAI-1 (plasminogen activator inhibitor-1), as being a direct target of TSP-1 and a node of convergence for different pathways that contribute to excessive secretion of VEGF by AVM-BECs. These pathways include those involving BMP4 (bone morphogenetic protein 4)/ALK (activin receptor-like kinase)2/ALK1/ALK5, Notch, and HIF-1α (hypoxia inducible factor 1α) signaling, all of which were upregulated in AVM-BECs. Two direct targets of miR-18a were identified, BMP4 and HIF-1α, and treatment of AVM-BECs with miR-18a mimic normalized activity of pathways associated with these targets, as well as secretion of VEGF by AVM-BECs. Important from a potential therapeutic standpoint, miR-18a did not alter pathway activity in non-AVM BECs. Interestingly, the effect of miR-18a on HIF-1α/VEGF signaling only occurred during normoxia, which is the normal state of the AVM nidus.⁷ The loss of regulation of HIF-1 α and VEGF secretion by miR-18a under hypoxic conditions further reveals context dependency of miR-18a’s antiangiogenic activity.⁸ Under normoxia, miR-18a can inhibit aberrant angiogenesis in AVM-BECs through simultaneous inhibition of BMP4, Notch and HIF-1α pathways, but the antiangiogenic activity of miR-18a is neutralized when the HIF-1α pathway is upregulated by hypoxia.

Successful delivery of synthetic miRNA molecules into target tissues remains one of the greatest challenges in the field of miRNA therapeutics.⁹ For targeting brain tissues, this issue is further complicated by the fact that only lipid-soluble small molecules less than 400 daltons are able to cross the blood brain barrier (BBB). Local delivery of miRNA therapeutics directly into the brain has been demonstrated to be one of the most relevant delivery routes, however, precisely targeting a specific brain region without damaging surrounding tissues can be challenging. Another approach to circumvent the lack of permeability of the BBB is temporary disruption of the BBB using chemical agents such as alcohols or vasodilators.¹⁰ In the current study, the authors’ strategy to target brain AVMs in vivo was to co-administer miR-18a, argonaute 2 (Ago-2) and NEO100 to Mgp (matrix GLA protein)^−/− mice intranasally (IN). NEO100 is a good manufacturing practices-quality perillyl alcohol, which, in pharmacokinetic studies, was shown to accelerate miR-18a delivery and enhance miR-18a abundance in brain tissues. Previously, this group had shown that co-administration of Ago-2 with miR-18a increased stability and cellular uptake of miR-18a.¹¹ High BMP4 activity in the Mgp^−/− model causes brain AVMs in 100% of the mice, and the authors used ultra high-resolution computed tomography angiography (CTA) to show that IN delivery of miR-18a improved the cerebrovascular of Mgp^−/− mice compared to untreated Mgp^−/− mice. Immunohistochemistry of brain sections from treated mice showed, as expected based on in vitro experiments, decreased expression of BMP4, PAI-1, VEGF, and HIF-1α. Interestingly, IN miR-18a also had a number of systemic effects: it normalized the number of bone marrow precursors and decreased polychromasia in Mgp^−/− mice, it reduced inflammation in the lungs and livers of these animals, and it appeared to restore functionality of the livers and spleens.

By demonstrating the ability of miR-18a to target multiple pathways relevant to AVM phenotype and the use of detailed, well-controlled brain imaging experiments, Marín-Ramos and co-authors have clearly demonstrated efficacy of a miRNA-based therapy in treating brain AVMs. Furthermore, the intranasal delivery system used to enhance uptake and bioavailability of therapeutics in the brain appears to have promise for miRNA delivery.¹² However, while the study provides some hope for patients who have inoperable AVMs or have a high risk of recurrence, clinical translation of this work is very preliminary. The model of brain AVM used in this study is severe, with life expectancy of two months because of disease in all major organs.¹³ The mice were treated with IN miR-18a for only two weeks due to problems with survival; most Mgp^−/− mice did not survive until the end of treatment. So, while IN miR-18a treatment improved cerebral vasculature and disease in other organs, it is not clear that this treatment improves outcomes. Further testing of the efficacy of IN miR-18a and its impact on survival in other models of brain AVM is warranted. Other models might also be valuable in providing insight into miR-18a-mediated modulation of pathways not identified by in vitro studies. For example, TGF-β (transforming growth factor beta) signaling has been shown to be involved in AVM pathogenesis,¹⁴ but this pathway was not affected by miR-18a treatment of AVM-BECs. Whether IN miR-18a can modulate TGF-β signaling during AVM development in other models of AVM is not clear.

Pharmacokinetic studies indicated that IN delivery of miR-18a was at least as efficient as intravenous delivery, while having the advantage of being less invasive. However, IN delivery of miR-18a delivery system clearly had systemic off-target effects, including effects on bone marrow. While the off-target effects appear to have positive implications for the overall health of the animals in this study, off-target effects, in general are a major hurdle for miRNA-based therapeutics. Furthermore, this finding brings into question the molecular basis underlying the effect of IN miR-18a on abnormal brain vasculature. Work by Lu and colleagues has indicated increased expression of endothelial progenitor cells (EPCs) in higher-stage AVMs,¹⁵ which raises the question of whether the actions of IN miR-18a on brain AVMs in Mgp^−/− mice may be due, at least in part, to off-target effects on bone marrow and mobilization of bone-marrow derived EPCs.

Recent years have witnessed significant progress in small RNA-based therapeutic strategies that are likely to significantly expand in the coming years. Currently, there are two FDA-approved small-interfering RNA (siRNA) drugs, patisiran (Onpattro; Alnylam) and givosiran (Givlaari, Alnylam), that target liver mRNAs for treatment of hereditary transthyretin amyloidosis and acute hepatic porphyria, respectively. Several other siRNA drugs are in development, including the PCSK9 (proprotein convertase subtilisn/kexin type 9)-targeted inclisiran (Alynlam), which has met all primary and secondary endpoints across three phase II trials and matched the LDL-lowering efficiency of antibody-based PCSK9 inhibitors. Although the emergence of miRNA-based therapeutics has not yet translated into FDA-approved drugs, there are several candidates that are in clinical development or have been tested in phase I and phase II trials, including MRG-110 (miRagen), an antisense oligonucleotide against miR-92; MRG-106 (cobomarsen, miRagen), an antisense oligonucleotide against miR-155; and MRG-201 (miRagen), a miRNA mimic of miR-92. The present study is an important contribution to research efforts to identify new candidate miRNA therapeutics and determine how these therapeutics could be used clinically.