Ribonucleic Acid (RNA) Therapeutics: Role of Long Noncoding RNAs in Ocular Vascular Diseases

The recent annotation of the complete sequence of human genome totals ∼64,000 genes, of which about 20,000 genes are predicted to be protein coding. ¹ This indicates that ∼69% of human genes are transcribed as noncoding ribonucleic acids (ncRNAs). Two major classes of ncRNAs include the well-studied microRNAs (miRNAs) and the recently identified long ncRNAs (lncRNAs), which are defined as ncRNAs longer than 200 nucleotides that do not encode proteins. Compared with mRNAs, lncRNAs tend to show lower expression, poorer sequence conservation, and more specific tissue expression patterns. LncRNAs are more often localized in the nucleus than in the cytoplasm and have diverse functional mechanisms.

In the nucleus, lncRNAs can function in cis or in trans to influence gene transcription by functioning as scaffold, decoy, guide, or signal to regulate chromatin remodeling. ² Cytoplasmic lncRNAs can regulate mRNA stability and translation by interacting with RNA-binding protein, acting as antisense RNA to regulate mRNA stability, acting as competitive endogenous RNAs (ceRNAs) or “sponges” for miRNAs, regulating protein translation and affecting post-translational mechanisms, or encoding micropeptides. Some lncRNAs are not linear but form covalently closed circles (circRNAs), which will be discussed specifically in future issues.

An increasing number of lncRNAs have been linked to ocular vascular diseases in patients or animal models. For example, lncRNAs ANRIL and MIAT1 expression is upregulated in the retina of diabetic or diabetic retinopathy (DR) animals, and in the serum/vitreous or plasma of DR patients.^3–8 HOTAIR, HIF1A-AS2, and SNHG16 expression is upregulated in the serum/plasma or vitreous of proliferative DR patients.^9–15 MALAT1 expression is upregulated in the oxygen-induced retinopathy mouse model and diabetic patients,^16–19 whereas Vax2os1 and Vax2os2 expression is upregulated in the aqueous humor of wet age-related macular degeneration (AMD) patients. ²⁰

In contrast, MEG3 expression is downregulated in DR serum and streptozotocin (STZ)-induced diabetic mice,^21–24 and PKNY expression is decreased in the retinal pigment epithelium/choroid in a laser-induced choroid neovascularization wet AMD model. ²⁵

Functionally, studies using animal models have established ANRIL, MIAT1, HOTAIR, and MALAT1 as proangiogenic lncRNAs, whereas MEG3 and PKNY as antiangiogenic lncRNAs. In addition, ANRIL is also proinflammatory and HOTAIR could regulate vascular permeability. Particularly, lncEGFL7OS is a human/primate specific lncRNA that enhances choroid sprouting angiogenesis in human choroid tissue ex vivo. ²⁶ These lncRNAs could have therapeutic implications in ocular vascular diseases, including wet AMD, proliferative DR, and retinopathy of prematurity.

RNA-based therapies have gained momentum due to the approval of RNA drugs by the FDA in recent years, including antisense oligonucleotides (ASOs), small interfering RNAs (siRNAs), and mRNA-based vaccine for coronavirus disease 2019 (Covid-19).^27,28 Not surprisingly, many of these technologies could be applied to lncRNAs for preclinical or clinical development. For example, ANRIL and HOTAIR were shown to be proangiogenic and proinflammatory, ASO or siRNAs to these lncRNAs could be tested in models of wet AMD or proliferative DR.

For antiangiogenic lncRNAs such as MEG3 and PKNY, their overexpression using adeno-associated viruses or lentiviruses could have therapeutic potential in aforementioned diseases. Similar to mRNA, lncRNAs are susceptible to degradation. However, circRNAs are highly stable as they are protected from exonuclease-mediated degradation. Recently, circRNA vaccines to SARS-CoV2 have been developed, which have shown favorable protection against SARS-CoV2 variants. ²⁹ Therefore, developing circ-lncRNA together with lipid nanoparticles delivery could be a potential option for lncRNA overexpression in disease models.

As lncRNAs have diverse function mechanisms, the mode of action of lncRNAs should also be considered when designing therapies. For lncRNAs that regulate gene transcription, targeting lncRNA-regulated genomic loci using CRISPR-Cas9, or lncRNA using Cas13, could offer therapeutic opportunities. For example, dCas9-KRAB-mediated targeting of the EGFL7/miR-126/lncEGFL7OS locus has been shown to inhibit the expression of genes in this locus and repress human angiogenic activities in vitro. ²⁶ For lncRNAs that provide docking sites or interactional structural elements for miRNAs, DNAs, or proteins, designing small molecules that could interrupt the functional domain of lncRNAs or affect their interaction with partner molecules could have therapeutic implications. For example, small molecules have been designed to target MALAT1 triple helix, which reduced the levels of MALAT1 and its downstream genes. ³⁰

The future is bright for lncRNA-based therapies in ocular vascular diseases. However, some challenges exist, including undercharacterization of the diverse functions of individual lncRNAs, instability and long-term efficacy of lncRNAs or siRNAs-based therapies, off-target effects of siRNAs, inefficient targeted delivery of negative charged RNAs, and the potential immunogenicity of long RNAs.