Red blood cells (RBCs) are actively involved in oxygen and carbon dioxide transportation, and their unstoppable journey confers transportability advantages. RBCs are known to have a small amount of miRs and mRNAs. In contrast to free nanoparticles, RBC-associated nanoparticles have been shown to deliver ten times more drugs to the target site. The half-life of RBC is 10–15 days in comparison to PEG liposomes (3–6 h), filomicelles (1–3 days), polymeric micelles (0.1–6 h), proteins and conjugates (10 min–6 h), and polymersomes (10–12 h). Due to an extended half-life compared to other artificial career systems, RBCs remain longer in the blood circulation, reaching all organs and tissues throughout the body. The delivery of therapeutics using RBCs could be achieved by encapsulating drug-loaded nanoparticles within or by attaching them to the surfaces of erythrocytes. Usman et al. demonstrated that RBC-derived extracellular vesicles deliver the miR-125b inhibitor and reduce cancer progression.
T cells are omnipresent in the body, though most are present in lymphoid tissues. These cells recognize, attack, and eradicate cancer cells and, thus, can act as a naturally occurring delivery vehicle for anticancer modalities. As these cells are nucleated, the abundance of nucleic acids might interact and dilute the efficacy of the miR/miR inhibitor. However, the miR/miR inhibitor encapsulation within a T cell could protect it from blood ribonuclease, minimize its interaction with the cellular DNA/RNA, and selectively deliver miRs at the target site. Recently, Parayath et al. reported that injectable nanocarriers providing in vitro transcribed chimeric antigen receptor mRNAs transiently reprogram circulating T cells to recognize disease-relevant antigens. They demonstrate that the infusion of these nanocarriers induces host T cells to express tumor-specific chimeric antigen receptors to cause disease regression as efficiently as achieved by the infusion of ex vivo engineered lymphocytes.
Dendritic cells (also known as antigen-presenting cells) play a crucial role in regulating immune responses. Firdessa-Fite et al. compared a mRNA-based dendritic cell delivery system against lipid nanoparticles targeting CD4+ and CD8+ T cells. The injection of both delivery systems via different routes resulted in mRNA delivery to the respective sites, instigating T-cell responses. These cells were explored for the delivery of miRs. The intravenous administration of a Cy5-tagged miR let-7-loaded extracellular vesicle conjugated with aptamer AS1411 decreased the tumor growth in vivo. Given the efficient delivery of oligonucleotides, dendritic cells hold potential for miR and miR inhibitor delivery.
Macrophages are monocyte-derived cells that actively phagocytose foreign cells, cancerous cells, dead cells, dying cells, and microbes. The tumor-targeting ability of macrophage-derived exosomes and their cell membranes can be harnessed as a tool for delivering miRs or miR inhibitors to tumors. The macrophages take up liposome-entrapped RNA molecules, some of which remain undegraded and are released via microvesicles. Wayne et al. demonstrated that macrophages loaded with calcium integrin-binding protein-1 (CIB1) siRNA result in decreased tumor growth and decreased mRNA expression of CIB1 and KI67 in MDA-MB-468 human breast cancer cells. Akao et al. demonstrated that the chemically modified miR-143 containing microvesicles was secreted from THP-1 macrophages following intravenous injection. These results suggest the suitability of using macrophages-derived vesicles as a delivery system for miR inhibitors. Besides, as macrophages are heavily involved in the pathogenesis of multiple diseases, understanding how to use macrophages for drug delivery can benefit the treatment efficiency.
Natural killer cells migrate to the inflammation site and attack cancer cells; therefore, they can be used for the drug delivery of anticancer miR modulators. Neviani et al. reported that the natural killer-derived exosome miR-186 inhibits neuroblastoma growth and immune escape mechanisms. Further, natural killer-derived exosomal vesicles inhibit the malignant transformation of pancreatic cancer, which requires the presence of miR-3607-3p.
3. Challenges in the Development of MiR Therapeutics
Less than twenty miR therapeutics have entered clinical trials, and none progressed to phase III trials. In contrast, over 60 siRNA drugs are in clinical trials, with two approvals. In a direct comparison between miR therapeutics and siRNA therapeutics, a higher percentage of terminated/suspended clinical studies (50% vs. 35%) and none in phase III clinical trials (0% vs. 12%) suggest unexplainable hurdles in the development of miR therapeutics. However, this could be because of the fewer miR therapeutic candidates in the pipeline. Exogenous siRNAs are designed explicitly against the target gene. On the other hand, endogenous miRs target multiple genes. Zhang et al. analyzed the targets of miRs and siRNAs that entered the clinical trials and found that miR targets ranged between 30 and 250 while that of siRNAs ranged between 1 and 3. The siRNAs exhibit 100% sequence complementarity with the target gene, whereas the sequence complementarity of endogenous miRs range between 20 and 90%. miRs target tens to hundreds of genes and inhibit gene expression by decreasing the stability or inhibiting the translation. Still, in specific conditions (e.g., starvation), they upregulate the expression of genes. The disease pathway analysis of a miR-34a mimic predicted the deregulation of two immune pathways, and it was discontinued in phase I clinical trial due to immune-related severe adverse effects. Nonetheless, the disease pathway analysis of miR-122 showed only a small number of genes linked with jaundice, and it did not appear in the phase I subjects but appeared in the few cases in phase II. Therefore, the multiple targets of miRs are a crucial challenge in the development of miR therapeutics. Although this is the inherent characteristic of miRs, this limitation can be debilitated by targeted delivery and chemical modifications to the molecules that improve their binding characteristics to the target genes. Further, the use of miRs derived from the plants that affect a gene that belongs to the pathogen (e.g., aly-miR396a-5p) and does not target genes in the host could be utilized to develop miR therapeutics.
The development of miR therapeutics would require the identification of biologically significant miR in a disease condition, efficient delivery of miR modulators to the target cell/tissue, regulation of target(s) expression, and minimal nonspecific interactions. The biological complexity and physiochemical nature of the miR/miR inhibitor pose challenges in developing miR therapeutics, as described below.
3.1. The Negative Charge of the MiR Inhibitors
Most DNA/RNA-based miR modulators are negatively charged, promoting their nonspecific binding to the blood proteins and decreasing urinary clearance. As an example, a phosphorothioate-modified single-stranded oligonucleotide, which is negatively charged, is quickly transmitted from the blood to tissues (minutes to hours) and exhibits a half-life of 2–4 weeks. In contrast, oligonucleotides that lack a charge (e.g., PNAs and unmodified siRNA) weakly bound to plasma proteins exhibit a rapid clearance either due to metabolism in the blood or excretion via urine, leading to a lower tissue uptake. Therefore, the nonspecific accumulation of ASOs in tissues could lead to sequence complementarity-dependent effects in nontarget tissues. The common concerns with the synthetic nucleic acid analogs include (a) a moderate efficacy and (b) adverse effects, such as prolonging the activated partial thromboplastin time, activating the complement cascade, and accumulating nonspecifically in tissues. The activated partial thromboplastin time depends on the plasma concentration of the nucleic acid analogs and is not considered clinically significant. However, the accumulation of nucleic acid analogs in tissues contributes to adverse outcomes that depend on the negative charge of these analogs and leads to nonspecific binding to proteins. The negative charge on miR modulators negatively affects the cellular entry of these molecules via passive diffusion as the cell membrane is also negatively charged.
3.2. Functional Complexity and Off-Target Effects
A single miR targets hundreds to thousands of mRNAs, and a single mRNA is targeted by tens to hundreds of miRs, making identifying one specific miR that controls a disease pathogenesis challenging. Although in silico tools can predict the binding strength of a miR to the target mRNA, the biological effect depends on multiple factors, such as expression of miR, availability of AGO2, the abundance of the target gene of interest, and expression of the miR target genes. Besides, the expression of target genes in other cell types and tissues can be targeted by miR modulators, which often contribute to unwanted effects. Most of the DNA-based miR modulators have a negative charge that leads to their binding with macromolecules, leading to nonspecific tissue accumulation and prolonged tissue half-lives (2–4 weeks). Therefore, identifying the miR that plays a therapeutically significant role in disease pathogenesis is a challenge.
3.3. Stability of MiR Modulators in the Blood and Endosomal Escape
miRs and unmodified miR inhibitors are prone to degradation by ribonuclease enzymes in a biological system. Naked miRs with an unmodified 2′ OH in ribose are quickly degraded in the blood by the serum RNase. Modifying 2′ OH (e.g., locked nucleic acid, phosphorothioate, O-(2-Methoxyethyl), O-Methyl, and Fluro) can improve the stability and affinity of the target genes. Ribonuclease enzymes that degrade miRs invade the duplex via the sense strand’s less stable terminus (3′ terminus). Chemical modifications have been developed to prevent this. Further, chemical modifications are not well-tolerated at the 5′ end of the antisense strand, highlighting how crucial molecular asymmetry plays in RNAi activity. This evidence suggests that phosphorothioate modifications at the sense strand prevent degradation and improve serum stability. The intracellular trafficking of miRs begins in the early endosome, which later merges with late endosomes and lysosomes with acidic environments where nucleases degrade them. For the miRs inhibitors to engage with miRs and the miR mimic to engage with miRISC, they will need to escape the endosome, which constitutes a challenge. Strategies such as pH-sensitive lipoplexes/polyplexes and photosensitive molecules can promote an endosomal escape.
3.4. Chemical Modifications and Cellular Entry of MiR Inhibitors
A vital issue in miRs therapeutics is the sequence-independent and complement-mediated toxicity owed to the chemical modifications (e.g., phosphodiester linkages, ribose backbone, and 2′-O-(2-methoxyethyl)) made in an attempt to mitigate the nuclease-mediated degradation of miR modulators. The inhibition of blood coagulation, complement cascade activation, immune cell activation, and reduced peripheral white blood cell counts are often observed as signs of toxicity arising from such chemical modifications. Phosphorothioate-based miR inhibitors exhibit poor uptake in the skeletal muscle and brain due to their hydrophilic and negatively charged nature and inability to cross tight junctions, such as the blood–brain barrier. However, these molecules are stable; exhibit a long half-life; and show excellent penetration of some tissues (e.g., kidney, liver, spleen, lymph nodes, and bone marrow). The sizes of miR inhibitor-loaded nanoparticles determine their ability to penetrate through the pores of leaky vessels. The pH affects the formation of smaller and monodispersed nanoparticles, while the outer surface hydrophilicity of nanoparticles determines their urinary clearance. Thus, in addition to optimizing the chemical nature of miR inhibitors, the optimization of chemical conditions during nanoparticle synthesis and modifications of the molecules on the outer surfaces of nanoparticles can affect the delivery of the miR inhibitors to the target tissues. While the various deliveries, as mentioned above, alongside cautious dosing, resolve this concern to some extent, further studies are warranted to generate biocompatible and efficient miR therapeutics.
4. Opportunities for the Development of MiR-Therapeutics
The development of more biocompatible and efficient miR inhibitors (e.g., γ-modified PNA-based miR inhibitors), the identification of plant-based miRs with biologically significant effects in animals by targeting the expression of a specific gene (e.g., aly-miR-396a-5p), and the microvesicle-dependent targeted delivery of miRs are emerging opportunities for the development of miR therapeutics.
4.1. Gamma-Modified PNA-Based MiR Inhibitors
Several synthetic nucleic acid analogs have been employed to target miRs. However, various challenges still need to be resolved for miR inhibitor-based therapeutic modalities to be safe and effective in patients. Peptide nucleic acids (PNAs) are neutral charge synthetic DNA/RNA mimics in which the phosphodiester backbone is substituted with an N
-(2-aminoethyl) glycine backbone. Due to poor water solubility, the regular PNAs did not progress as the chemistry of choice. However, a novel class of PNAs called gamma PNAs (γPNAs) are highly water-soluble and charge-neutral; they neither aggregate nor adhere to surfaces or other macromolecules in a nonspecific manner. As individual strands, γPNAs adopt a right-handed helical motif—as confirmed by circular dichroism, nuclear magnetic resonance, and X-ray crystallography—and hybridize to DNA or RNA with an unusually high affinity and sequence specificity 54. Prior studies revealed that, on average, each γ modification stabilizes a PNA–RNA duplex by 5 °C. γPNA is the only molecule known to invade duplex RNA/DNA composed of up to 100% CG content in a sequence-specific manner 
. Another attractive γPNA design is tail-clamp modifications, which can be used to target miRNA-containing homopurine stretches.
Cheng et al. demonstrated that the γPNA-based miR-155 inhibitor attached to a pH-induced transmembrane structure produces a novel construct that targets the tumor and inhibits its growth in vivo 
. In another study, the pre-radiation administration of γPNA-based antisense oligonucleotide against double-stranded break repair factor KU80 significantly increased the antitumor effects of radiation without showing any evidence of toxicity. In two additional studies, γPNA-based miR inhibition showed efficacy against ischemia-reperfusion injury and cancer in vivo. Gupta et al. showed that γPNA delivered via polymeric nanoparticles antagonized the expression of miR-210 in a HeLa cancer model, resulting in an anticancer effect with no observed toxicities.
γPNAs have been exploited in several biological and biomedical applications (e.g., electronic barcoding, nanotechnologies, gene editing, and gene targeting). A comprehensive cytokine and metabolic analysis supported the in vivo biocompatibility of γPNAs. Compared to DNA-based miR inhibitors, PNAs are rapidly eliminated via urinary clearance (blood half-life: hours vs. weeks), possibly because of their minimal interaction with the blood proteins. Beavers et al. studied the blood pharmacokinetics of the intravenously injected PNA (non-gamma-modified)-based miR-122 inhibitor and found that the blood half-life was 0.9 min, which was increased to >30 min when loaded onto porous silicon nanoparticles. They also measured the biodistribution of the Cy5-labeled PNA-based miR-122 inhibitor alone and loaded onto porous silicon nanoparticles 160 min after the intravenous administration, and found the presence of an inhibitor in multiple tissues (e.g., heart, lung, liver, and kidney). The lower retention may require a higher dosing frequency to achieve the desired miR inhibition and constitutes a limitation. However, it is outweighed by the exceptional affinity, specificity, and biocompatibility of γPNA-based miR inhibitors. Therefore, employing γPNA technology represents a novel approach to developing miR therapeutics.
4.2. Plant-Derived MiRs
Plant-derived exosomes contain miRs and can be taken up by the microbes residing in the gut or intestinal epithelial cells. The presence of the hyphen (-) symbol in animal miR distinguishes them from plant miR (e.g., miR-159 vs. miR159). Intestinal microbes take up plant-derived exosomes, and the lipid compositions of exosomes determine their selection by the microbes. For example, ginger exosome-like nanoparticles containing mdo-miR7267-3p are preferentially taken up by Lactobacillaceae. mdo-miR7267-3p targets monooxygenase ycnE, which leads to increased levels of indole-3-carboxaldehyde and IL-22. This study showed that ginger exosome-like nanoparticle RNAs ameliorate mouse colitis via IL-22-dependent mechanisms. Several studies have reported that plant-derived miRs are stable and cross the gastrointestinal tract (e.g., miR156a, miR168a, and miR159a/e). miR168a derived from Moringa olifera seeds was predicted to be a human analog of miR-579. Its transfection in the hepatocarcinoma cell line induced a significant decrease in the expression of SIRT1, an experimentally verified target of miR-579. Plant miR2911 shows better stability and gastrointestinal tract absorption than synthetic miR2911, which could be attributed to its association with plant proteins or plant-specific 2′-O-methylation. Mlotshwa et al. tested the antitumor effects of an orally delivered cocktail of 2′-O-methyl-modified antitumor miRs (containing miR-34a, miR-143, and miR-145) and demonstrated a significant decrease in the tumor burden in mice. The effect of ginger-derived miRs (aly-miR396a-5p and rlcv-miR-rL1-28-3p) depends on the specific inhibition of the spike protein and nonstructural protein 12 expressions . In another study, the miR containing microvesicles isolated from Moringa oleifera (Drumstick) were bioinformatically predicted to target apoptosis-related human genes, and its exposure decreased the cancer cell viability and increased apoptosis. From rice, multiple miRs (e.g., osa-miR156-5p and osa-miR164-5p) are predicted to target human genes. Chin et al. reported the presence of plant miR159 in human serum and showed that its levels inversely correlated with breast cancer incidence and progression. Further, they showed that miR159 targets TCF7, which is highly expressed in cancerous cells and effectively reduces the tumor volume in mice. Thus, miR159 could potentially be used as a therapeutic agent against breast cancer. The precise uptake mechanism for plant-derived miR is not well-known, but miR-loaded exosome-like nanoparticle uptake by intestinal epithelial stem cells by micropinocytosis has been shown. Further, stomach pit cells absorb dietary and orally administered miRs in a SIDT1-dependent manner. Once internalized by the intestinal epithelial cells, they can be packaged into microvesicles and subsequently released into the bloodstream.
Although these reports greatly support the potential of plant-derived oral-administered miRs as cross-kingdom regulators, data reproducibility and technical errors are still a reason of concern. Dickinson et al. compared small RNA sequences in the serum and liver of mice fed a rice-based chow (up to 75% rice) with that of mice fed normal chow and found <10 reads per 10 million reads for the plant-derived miRs. In another study, feeding mice with a diet supplemented with corn miRs for two weeks was not associated with a significant increase in their levels in the blood or tissues, suggesting the extensive degradation of miRs during the digestive process . Further, the presence of diet-derived plant miRs at low levels and their presence in only a few animals (out of all that received a high corn diet) suggests that a leaky gut in mice caused by pathological conditions (e.g., cancer and inflammation) could contribute to the entry of orally administered miRs. The gastrointestinal uptake of miRs increased when mice were provided with aspirin or anti-CD3 antibodies, which increase gut permeability. A comprehensive analysis of the sequencing dataset from human serum showed the presence of plant miRs at low levels, and no correlation between their levels in the tissues that were exposed to dietary miRs (liver) and that were not (cerebral spinal fluid) suggested the possibility of contamination. However, the low reads for plant-derived miRs in the study by Dickinson et al. could be because of the sequencing bias of the approach for plant and animal miRs. As, unlike most of the previous high-throughput studies that detected the most abundant miRs (miR156 and miR168 of at least 10,000 reads per million), this study detected no miR-156 and shallow levels of miR168 (<200 per million) . The plant-derived miRs have recently been actively discussed in the scientific community to understand their potential and loopholes as miR therapeutics . We believe that plant-derived miRs are scalable, and pending validation by independent studies, they represent a unique opportunity to be developed as modulators of gene expression with therapeutic potential.
4.3. Cell-Targeted Delivery Using Microvesicles and Exosomes
Recent studies have shown the effectiveness of cell-derived microvesicles or exosomes in delivering oligonucleotides. Several studies have shown that the transfer of miRs (e.g., miR-302b, miR-142-3p, and miR-223) by platelet-derived microvesicles affects the endothelial gene expression and inflammation of cells. Greater than fifty percent of peripheral blood microvesicles are likely derived from non-nucleated platelets and have been shown to deliver miRs effectively. The experimental approach where platelet-derived microvesicles are loaded with the oligonucleotide of interest and modified to target a specific cell type, in our opinion, holds high promise as a delivery vehicle.