Unlocking the potential of chemically modified peptide nucleic acids for RNA-based therapeutics

Abstract

RNA therapeutics have emerged as next-generation therapy for the treatment of many diseases. Unlike small molecules, RNA targeted drugs are not limited by the availability of binding pockets on the protein, but rather utilize Watson–Crick (WC) base-pairing rules to recognize the target RNA and modulate gene expression. Antisense oligonucleotides (ASOs) present a powerful therapeutic approach to treat disorders triggered by genetic alterations. ASOs recognize the cognate site on the target RNA to alter gene expression. Nine single-stranded ASOs have been approved for clinical use and several candidates are in late-stage clinical trials for both rare and common diseases. Several chemical modifications, including phosphorothioates, locked nucleic acid, phosphorodiamidate, morpholino, and peptide nucleic acids (PNAs), have been investigated for efficient RNA targeting. PNAs are synthetic DNA mimics where the deoxyribose phosphate backbone is replaced by N-(2-aminoethyl)-glycine units. The neutral pseudopeptide backbone of PNAs contributes to enhanced binding affinity and high biological stability. PNAs hybridize with the complementary site in the target RNA and act by a steric hindrance­–based mechanism. In the last three decades, various PNA designs, chemical modifications, and delivery strategies have been explored to demonstrate their potential as an effective and safe RNA-targeting platform. This review covers the advances in PNA-mediated targeting of coding and noncoding RNAs for a myriad of therapeutic applications.

INTRODUCTION

Dysregulation of gene expression in disease can be caused by various genetic aberrations, ranging from single base-pair mutations to multibase rearrangements (Alexandrov et al. 2013). The Human Genome Project (HGP), led by the International Human Genome Sequencing Consortium (IHGSC), published the first human genome sequence (complete to 92%) in 2001 (Venter et al. 2001). In the last two decades, information obtained from the HGP has been crucial in advancing our understanding of basic biology, mutations driving disease pathogenesis, and identifying new drug targets. Recently, the remaining 8% of the human genome covering the chromosomal gap assemblies except for the Y chromosome was reported (Nurk et al. 2022). HGP established that only ∼1.5% of the genome accounts for coding genes, while more than ∼98% is noncoding in nature (Nurk et al. 2022). Over the last few years, several classes of noncoding RNAs, including microRNA (miRNA), short interfering RNA (siRNA), long ncRNA (lncRNA), circular RNAs (circRNA), piwi-interacting RNA (piRNA), and enhancer RNA (eRNA), have been identified (Palazzo and Lee 2015). These noncoding RNAs play an essential role in regulating gene expression. Moreover, mutations in both coding and noncoding genes have been identified as the root cause of many disorders (Saliminejad et al. 2019). The HGP established genome-wide association studies correlating individual genetic variants with diseases (Hood and Rowen 2013). With the advancement in sequencing and bioinformatic analysis, the ease of sequencing individual genomes and identifying their disease variants has provided new avenues to treat and possibly cure diseases.

The aforementioned developments have been the launchpad for precision or personalized medicine, where complementary treatments can be developed, and diseases can be targeted, after confirming the aberration in the RNA and DNA sequences. Hence, various synthetic chemically modified nucleic acid–based modalities have emerged to target RNA and DNA sequences by Watson–Crick (WC) base-pairing rules and repair misregulated gene expression. In recent years, nucleic acid and antisense oligonucleotide-based drugs that target aberrant RNAs, from companies like Ionis, Alnylam, and Sarepta, have gained approval from the United States Food and Drug Administration (FDA) for treating rare genetic disorders (Crooke et al. 2021).

In particular, RNA interference (RNAi)-based double-stranded siRNAs have been widely used to efficiently target the mRNAs in a sequence-specific manner. siRNAs associate with the Argonaute enzymes (AGO) to form the RNA-induced silencing complex (RISC) and induce AGO-mediated degradation of the target mRNA. The in vivo delivery of highly charged and double-stranded siRNAs is achieved using lipid-based nanoparticles (LNPs) or enhanced stabilizing chemistry (ESC). Patisiran (Onpattro), the first siRNA drug approved by the FDA, was formulated as LNPs for the treatment of hereditary transthyretin-mediated amyloidosis (hATTR). ESC-based strategies utilize conjugates like N-acetyl galactosamine (GalNAc), which recognizes the asialoglycoprotein receptor (ASGPR) on hepatocytes to achieve targeted liver delivery with minimal immune stimulation. Recent siRNA products, including givosiran, lumasiran, and inclisiran, are approved as GalNAc conjugates for the treatment of hepatic porphyria, primary hyperoxaluria type 1, and heterozygous familial hypercholesterolemia, respectively (Hu et al. 2020).

Besides siRNA, single-stranded antisense oligonucleotides (ASOs) represent another class of drugs which is widely used to target coding and noncoding RNAs. ASOs recognize their target via WC base-pairing rules and act either by activating RNase cleavage or steric hindrance–based mechanisms. RNase cleavage is mediated by the ribonuclease H1 (RNase H1) enzyme. RNase H1 recognizes the DNA (ASO): RNA heteroduplex and catalyzes the hydrolytic degradation of the target RNA (Dhuri et al. 2020). Chemically modified ASOs including phosphorothioates (PS) in their backbone activate RNase H1–based cleavage. PS modified-ASOs contain a sulfur atom in place of nonbridging oxygen in the phosphodiester backbone which makes them more resistant to nuclease degradation. Fomivirsen, a fully PS-modified 21 mer ASO, was the first antisense drug approved by the FDA as an intravitreal injection for the treatment of cytomegalovirus (CMV) retinitis in 1998 (Perry and Balfour 1999). However, it was withdrawn from the market in 2001 due to the success of antiviral drugs. Recently approved ASOs, including mipomersen (Hair et al. 2013) and inotersen (Gales 2019), are designed as gapmers, which contain a PS-modified central region and 2′ ribose modification (2′-O-methoxyethyl) on the termini. Gapmer ASOs activate RNase H1 and exhibit increased potency and metabolic stability (Crooke et al. 1995). Currently, there are several new ASO chemistries being evaluated in different stages of clinical trials for the nervous system, muscle, cardiovascular, metabolic, eye, and lung disorders, in addition to infectious diseases and cancer therapy (Crooke et al. 2021).

Locked nucleic acid (LNA), another chemical modification used in ASOs, is also being tested in clinical trials for antisense applications. LNA contains a ribose modification where the 2′ oxygen and 4′ carbon are constrained by a methylene bridge, resulting in increased binding affinity to the target RNA (Kumar et al. 1998). LNAs are designed as gapmers with a central DNA region surrounded by LNA nucleotides at the termini which stimulate RNase H1 cleavage. Further, ASOs designed as mixmers with alternating DNA and LNA stretches rely on a steric hindrance–based mechanism (Dhuri et al. 2020). LNA drugs under clinical trial are OGX-427 for metastatic bladder cancer and cobomarsen for the treatment of cutaneous T cell lymphoma (Gebert et al. 2014; Anastasiadou et al. 2021).

Chemically modified ASOs can also act by steric hindrance. The FDA-approved ASOs utilizing steric hindrance either contain ribose 2′-O-methyl modification or are phosphorodiamidate morpholino oligomers (PMO) (Aartsma-Rus and van Ommen 2007). In PMOs, the sugar is replaced with a morpholino ring, and the backbone contains a neutral phosphorodiamidate linkage. Nusinersen, approved for the treatment of spinal muscular atrophy (SMA), contains uniform 2′-O-methoxyethyl (2′-O-MOE) and PS modification (Havens and Hastings 2016). SMA is a neuromuscular disorder caused by loss-of-function mutations in the spinal muscular neuron 1 (SMN1) gene. SMN2, a homolog of SMN1, undergoes exon 7 exclusion during splicing to produce a truncated SMN protein. Nusinersen is designed to bind to the SMN2 pre-mRNA to alter the splicing and induce exon 7 inclusion in SMN2 mRNA which translates into the functional SMN protein (Havens and Hastings 2016). There are four PMO-based drugs, eteplirsen, golodirsen, vitolarsen, and casimersen, approved for the treatment of Duchenne muscular dystrophy (DMD). These PMOs modulate splicing by exon skipping–based mechanisms, where they bind to pre-mRNA and prevent the spliceosome from accessing the transcript splice site (Nan and Zhang 2018). Exon skipping by steric hindrance was also demonstrated with 2′-O-Me ASOs to target the β-globin pre-mRNA and modulate splicing in β-thalassemia (Aartsma-Rus and van Ommen 2007). Milasen is an N-of-1 ASO approved by the FDA to treat one patient, Mila, a 6-yr-old child suffering from Batten disease. Mila had a rare translocation mutation in the ceroid lipofuscinosis type 7 (CLN7) gene which resulted in mis-splicing and the introduction of a premature termination codon. The development of the milasen ASO was modeled after nusinersen in terms of chemistry, route of administration and initial dosing to induce exon skipping. Administration of milasen was able to reduce the frequency and duration of the seizures during the course of the treatment. Milasen's approval has shown ASO-mediated therapy's relevance in reducing the suffering of patients with rare disorders (Kim et al. 2019).

Peptide nucleic acids

Discovered by Nielsen and colleagues in 1991, peptide nucleic acids (PNAs) represent another class of nucleic acid analog which has been extensively explored to target RNA. Unlike other ASOs, PNAs contain a neutral pseudopeptide backbone of repeating N-(2-aminoethyl)-glycine units. Nucleobases (adenine, guanine, cytosine, and thymine) are attached to the backbone via a carboxy methylene linkage (Nielsen et al. 1991). The unnatural backbone of PNAs imparts enzymatic resistance against both peptidase and nuclease-mediated degradation (Demidov et al. 1994). Moreover, the chemical design of PNA ensures a similar distance between nucleobases to allow efficient hybridization with the complementary targets (DNA or RNA) via WC base-pairing. PNAs exhibit high binding affinity due to the reduced repulsion between the neutral PNA backbone and the negatively charged phosphodiester backbone of DNA or RNA (Nielsen et al. 1991). Thermal melting experiments established the order of duplex stability as PNA–PNA > PNA–RNA > PNA–DNA > DNA–DNA (Jensen et al. 1997). A PNA–RNA heteroduplex was found to be more stable and sensitive to mismatches than PNA–DNA. A single base-pair (bp) mismatch in a PNA–RNA heteroduplex reduced the thermal melting temperature by ∼12°C (Jensen et al. 1997).

Although the original PNA design was based on a neutral and achiral backbone, several chemical modifications have been explored to improve their hybridization properties. Multiple studies investigated the impact of introducing chemical moieties in the N-(2-aminoethyl)-glycine backbone to develop conformationally constrained PNAs with superior invasion kinetics. One of the earliest modifications was the substitution of D-lysine at the α-position (αPNA) of the PNA backbone. Thermal melting studies showed superior stability of D-Lysine αPNA: DNA duplexes relative to the regular PNA (Sforza et al. 2000). However, the introduction of a methyl group at the β-position (βPNA) showed comparable stability to unmodified PNA. Circular dichroism (CD) spectra of βPNA indicated helical preorganization, but this secondary structure did not improve the binding due to unfavorable steric interactions arising from β-methyl groups (Sugiyama et al. 2011). In 2007, Danith Ly's laboratory reported that γ-backbone modification (γPNA) can transform a randomly folded PNA into a helical structure (Fig. 1). They showed that incorporation of an L-serine side chain at the γ-position induces the preorganization of PNA into a right-handed helix and improved the hybridization stability with the target. A single L-serine γ-substitution in a 10-mer PNA increased the stability of a γPNA–RNA heteroduplex by ∼3°C. In addition, γPNAs showed better solubility and sequence specificity as a single mismatch lowered the Tm by 12°C–18°C for the target RNA (Dragulescu-Andrasi et al. 2006).

FIGURE 1.

Chemical structures of unmodified, αPNA, βPNA, and γPNA. The commonly utilized γPNA modifications include L-serine, miniPEG, and guanidine γPNA.

PNAs targeting mRNAs

Unlike gapmer ASOs, PNA–RNA hybrids are not substrates for RNase H1 enzyme. Hence, PNAs act via occupancy or steric hindrance–based mechanisms. The PNA–RNA hybrids are stable in serum and can be used to target cytosolic mRNA to inhibit translation. However, PNA-mediated translation inhibition requires the identification of sites sensitive to the translational arrest (Doyle et al. 2001). An initial study targeting the start codon (AUG) showed translation arrest of the protein in a dose-response manner in rabbit reticulocyte lysate (Koppelhus et al. 2002). It established the proof of concept, but in vitro validation was still required. Subsequent studies investigated different mRNA sites to define the rules for inhibiting their translation. PNAs targeting the 5′ untranslated region (UTR) and the translation start site of mRNAs were found to be more efficient in inhibiting their translation. These observations indicated that PNAs can block the formation of the ribosomal complex on the target mRNAs to inhibit translation initiation. One study reported a 30% decrease in mouse double minute 2 homolog (MDM2) protein using a 12 mer PNA sequence targeting the translation start site of MDM2 post-transfection in choriocarcinoma cells. Among the different PNA lengths (8 to 18 mer), a 12 mer PNA showed an optimum reduction in MDM2 and an increase in p53 protein levels. However, off-target effects of the 12 mer PNA were not evaluated (Shiraishi and Nielsen 2004). Further, PNAs targeting the coding region (CDS) of mRNA can also arrest translation by interrupting ribosomal elongation (Cheng et al. 2015).

Delivery of PNA targeting mRNA

The therapeutic utility of PNAs was stymied for decades due to poor cellular permeability and lack of efficient delivery strategies. With the advancement of diverse delivery platforms, these antisense applications of PNAs have expanded both in vitro and in vivo. To overcome the cellular delivery barrier, one study utilized poly (lactide-co-glycolic acid; PLGA)-based nanoparticles (NPs) to encapsulate the PNA targeting the translation start site of chemokine receptor 5 (CCR5) mRNA, up-regulated in HIV infections. The treatment of THP1 myelomonocytic cells with 3 mg of anti-CCR5 PNA loaded PLGA NPs resulted in a 50% knockdown of CCR5 mRNA (Bahal et al. 2013). PNAs have also been used to target the mutant variant of ataxin-3 (ATX3) mRNA, which causes Machado-Joseph disease. The ATX3 gene contains a tract of CAG repeats which leads to protein misfolding. Fibroblast cells treated with anti-CAG PNAs showed a ∼2.7-fold reduction in mutant ATX3 protein (Hu et al. 2011). The incorporation of a guanidine functional group at the γ position of PNAs increases their cellular uptake and results in the formation of highly stable duplexes (Sahu et al. 2009). Hence, gamma guanidine-PNAs (γ-GPNA) successfully inhibited the translation of oncogenic β-catenin, mutated in 20%–30% of hepatocellular carcinomas (HCC). HCC cells harboring the β-catenin mutation (HepG2 and SNU-499) showed a decrease in protein and cell viability (∼20%) upon treatment with γ-GPNA at only 1 µM dose (Delgado et al. 2013). Recently, PNAs have been used to target the surface glycoprotein CD5 which is involved in the progression of B-cell chronic lymphocytic leukemia (B-CLL). Peripheral blood mononuclear cells from B-CLL patients treated with PNA showed an ∼60% reduction in CD5 mRNA levels (Cesaro et al. 2022). PNAs have also been used to block miRNA binding sites on the 3′UTR of mRNAs, resulting in stabilization and increased protein production. This strategy has been used in stabilizing the cystic fibrosis transmembrane conductance regulator (CFTR) mRNA for the treatment of cystic fibrosis. PNAs targeting the binding site of miR-145-5p on CFTR mRNA led to a six- to eightfold increase in CFTR protein extracted from Calu-3 cells (Sultan et al. 2020). Multiple studies have established the potential of PNAs to efficiently inhibit and stabilize mRNA in vitro, but in vivo translation of these observations still needs to be validated. Recent advancements in delivery strategies for antisense PNAs like nanoparticles, ligands or peptide conjugates, have opened new avenues to evaluate their efficacy and safety in vivo.

PNAs targeting pre-mRNAs or splice-switching PNAs

Eukaryotic mRNAs are transcribed in the nucleus as pre-mRNAs containing coding exons interspersed with noncoding introns. The pre-mRNA splicing process ensures the precise and accurate excision of introns and ligation of exons to form the final mRNA, which can be translated into proteins. The spliceosome complex recognizes conserved splice sites at the intron-exon borders in pre-mRNAs to remove the introns. Many genetic diseases result from mutations in the splice sites which disrupt normal splicing and produce defective proteins (Blijlevens et al. 2021). Similar to the ASOs approved for DMD, antisense PNAs can be designed to block splicing from aberrant splice sites to induce exon skipping and restore the mRNA reading frame (Havens and Hastings 2016). A few studies have demonstrated the efficacy of PNA as splice-switching ASOs. Preliminary studies established the efficacy of a splice-switching PNA in an enhanced green fluorescent protein (EGFP) transgenic mouse model. Transgenic mice had an EGFR gene interrupted by an aberrant splice site containing a mutated intron of the human β-globin gene. Treatment of mice with PNA targeting the aberrant splice site restored the production of GFP in all organs (Sazani et al. 2002).

A splice-switching strategy has been clinically successful for the treatment of DMD, a neuromuscular disorder. DMD is caused by frameshift mutations in the dystrophin gene that produce truncated and nonfunctional proteins. Similar to other ASOs, a splice-switching PNA conjugated with Pip2b, a cell-penetrating peptide (CPP), efficiently targeted the mutation on exon 23 and increased the production of functional dystrophin in mouse mdx muscle cells (Ivanova et al. 2008). Another study designed a PNA to target the exon 6 splice junction of CD40 which plays an important role in immune responses. CD40 interacts with CD154 and triggers the activation of B and T lymphocytes, which produce a proinflammatory immune response in autoimmune diseases like rheumatoid arthritis (Lai et al. 2019). The B-cell lymphocytes treated with anti-CD40 PNA showed a reduction in CD40 and its downstream target, IL-12 (Siwkowski et al. 2004). PNAs inducing splicing alterations can also be used to inhibit the production of proteins which promote tumor progression. A PNA-based exon-skipping strategy was successfully used in Her-2-positive breast cancer. Her-2 is transcribed from the erbB-2 gene where exon 19 codes for the functionally active ATP catalytic domain. A PNA targeting the 5′ and 3′ splice site of exon 19 increased the exon 19 skipped fragments which produced inactive Her-2 in treated SK-BR-3 breast cancer cells (Pankratova et al. 2009).

AntimiR PNAs

Recent studies have highlighted the therapeutic application of PNAs as highly potent miRNA antagonists. miRNAs are 19–25 nt long noncoding RNAs which post-transcriptionally regulate the expression of multiple mRNAs (Saliminejad et al. 2019). miRNAs are transcribed and processed by the DROSHA complex in the nucleus to form pre-miRNAs which are transported to the cytoplasm. DICER cleaves the hairpin loop of pre-miRNA to form a functionally active miRNA duplex. Mature miRNA associates with argonaute (AGO) enzymes to form a miRNA-RNA-induced silencing complex (miR-RISC) (Saliminejad et al. 2019). miRNA guides the RISC assembly to the 3′-UTR of mRNA containing the complementary target site, a highly conserved “seed region.” Based on the extent of complementarity between miRNA and the target site, RISC assembly either induces translation arrest or AGO-mediated degradation of mRNA. miRNAs have been reported to regulate the expression of ∼10%–40% of protein coding genes; hence, dysregulation of miRNAs is associated with several pathologies (Saliminejad et al. 2019).

Genetic aberrations associated with cancer result in the up-regulation of miRNAs, called oncomiRs, and drive proliferation, angiogenesis, metastasis, and chemoresistance (Saliminejad et al. 2019). Hence, targeting oncomiRs has emerged as a novel anticancer intervention. PNAs designed to target the oncomiRs inhibit their functional activity by sterically blocking the interaction of miRNA with the target site (Cheng et al. 2015). Initial studies showed that a CPP conjugated PNA targeting miR-122 successfully inhibited the growth of hepatocellular carcinoma (Huh7) cells (Fabani and Gait 2008). Penetratin CPP containing six arginine on the N-terminus (R₆-Penetratin) assisted in the cellular delivery of PNA, and miR-122 inhibition was confirmed via northern blot analysis (Fabani and Gait 2008). OncomiR-21 has been reported to be elevated in both glioblastoma and prostate cancer. Multiple studies have reported the efficacy of antimiR PNAs targeting oncomiR-21 (Seo et al. 2019; Kim et al. 2020). AntimiR-21 PNAs showed an ∼50% decrease in viability of DU145 prostate cancer cells and reduced tumor metastasis in a xenograft mouse model (Kim et al. 2020). In an orthotopic mouse model of glioblastoma, antimiR-21 PNA containing NPs when delivered via convection enhanced delivery (CED) resulted in a >100% increase in median survival (Seo et al. 2019). PNAs were encapsulated in hydrophobic poly (lactic acid)-hyperbranched glycerol (PLA-HPG)-based polymeric NPs to achieve superior transfection and higher retention after intracerebral delivery. In a recent study, expression of miR-21 in tumor-associated macrophages (TAMs) was reported to promote tumor growth in a syngeneic mouse model of lung carcinoma. Targeted delivery of anti-miR-21 PNA using a pH low insertion peptide (pHLIP) to TAMs, reduced the tumor burden via increasing proinflammatory angiostatic function, decreasing vascularization, and inducing an antitumoral immune response (Sahraei et al. 2019).

miR-210 is up-regulated in solid tumors, including pancreatic, breast, colorectal, and cervical cancer (Gupta et al. 2017). Erythroid cells (K562) treated with anti-miR-210 PNAs showed a 50% decrease in miR-210 levels. Another study utilized miniPEG-γ-modified PNA loaded in PLGA NPs to target oncomiR-210 (Fabbri et al. 2011). γPNA NPs significantly reduced the growth of HeLa-derived xenografts after intratumoral delivery (Gupta et al. 2017). miR-155 is one of the well-established oncomiRs whose levels are up-regulated in several lymphomas, leukemias, and solid tumors like breast, pancreatic, lung, liver, and prostate. In a transgenic mouse model (mir-155^LSLtTA), overexpression of miR-155-induced lymphomagenesis, hence demonstrating the oncogenic role of miR-155. pHLIP-mediated systemic delivery of PNA targeting oncomiR-155 substantially improved the survival of a miR-155 addicted subcutaneous lymphoma mouse model without any toxic response. Further, pHLIP-PNA inhibited the oncomiR-155 activity in enlarged lymph nodes of mir-155^LSLtTA mice, with reduced liver and spleen metastasis (Cheng et al. 2015). In a recent study, we also designed tail clamp γPNA (tcγPNA-155), which can bind to the target oncomiR-155 via WC and Hoogsteen base-pairing to form a highly stable clamp and inhibit their activity. When delivered intratumorally, tcγPNA-155 inhibited tumor growth in xenograft mouse models of diffuse large B-cell lymphoma (DLBCL) (Dhuri et al. 2021a).

As an alternative to PNAs targeting the full-length of oncomiRs, we successfully established the efficacy and safety of short-cationic PNAs targeting the seed-region as a next-generation antimiR agent. Antiseed cationic 8-mer PNAs (containing three lysine or arginine residues) targeting the seed region of oncomiR-155 were designed and encapsulated in PLGA NPs to improve cellular delivery. Arginine containing antiseed PNA-155 showed superior inhibition of the target, and systemic delivery of PNA-155 loaded PLGA NPs prevented the growth of DLBCL-derived xenograft tumors (Malik et al. 2020). To further improve the delivery of PNA, we also developed positively charged PLGA/poly-L-histidine–based NPs to encapsulate the PNAs (Wahane et al. 2021). PLGA/poly-L-histidine NPs containing regular PNA-155 targeting miR-155 showed an approximately sixfold reduction in the growth of DLBCL xenografts. Importantly, multiple systemic dosing of PLGA/poly-L-histidine NPs did not induce any toxic response, hence establishing the safety of the delivery platform (Fig. 2).

FIGURE 2.

Graphical representation of various PNA design and delivery platforms explored for targeting miRs. (pHLIP) pH low insertion peptide, (PLGA NPs) poly(lactic-co-glycolic acid) nanoparticles, (PLA-HPG NPs) poly(lactic acid)-hyperbranched polyglycerol nanoparticles, (PLA-HPG-CHO) aldehyde functionalized poly(lactic acid)-hyperbranched polyglycerol nanoparticles.

In another study, we developed amphiphilic PNA constructs to improve their cellular delivery (Malik et al. 2022). Conjugation of lauric acid (C12) to the N terminus of γPNA (8 mer) targeting the miR-155 seed region imparted amphiphilic properties, resulting in their self-assembly. Further, formulation of γPNA-155-LA via ethanol injection resulted in ellipsoidal vesicles of ∼100 nm, exhibiting high cellular transfection and miR-155 inhibition in vitro. Transcriptome and proteomic sequencing indicated negligible off-target effects of seed targeting tiny LNAs on mRNAs containing the perfect match site (Obad et al. 2011). Hence, targeting the seed region of oncomiRs presents a viable approach to inhibit their functional activity.

The up-regulation of multiple oncomiRs is associated with reduced survival and poor prognosis in cancer patients. Hence, simultaneous inhibition of multiple oncomiRs has been explored to improve their anticancer efficacy. Prior studies have established the role of oncomiRs-155 and -21 in driving tumorigenesis in B-cell lymphoma. Hence, we utilized antimiR PNAs delivered via PLGA NPs to inhibit both oncomiRs-155 and -21 together, which showed an ∼80% reduction in viability of lymphoma cells compared to the single treatments (∼50%) (Milani et al. 2019; Dhuri et al. 2022). We also evaluated the antisense activity of PS- and PNA-based oligomers to inhibit multiple oncomiRs. Both PS and PNA ASOs, when delivered via PLGA NPs, showed comparable inhibition of miR-155 (∼90%) and miR-21 (∼50%) in lymphoma cells (Dhuri et al. 2022). Glioblastoma multiforme (GBM) is one of the most aggressive brain tumors with high recurrence rate and dismal survival. Several oncomiRs, including miR-10b, miR-21, miR-221, and miR-93, have been reported to drive the progression of GBM. Simultaneous inhibition of oncomiRs-155 and -221 combined with temozolomide (TMZ) was reported to induce significant apoptosis in the T98G glioma cell line (Milani et al. 2019). Targeting individual oncomiRs-10b or -21 has also shown anticancer efficacy in in vivo models of GBM. However, few studies have evaluated the coinhibition of oncomiRs-10b and -21 in GBM. In a recent study, we designed short (8-mer) cationic γ-modified PNAs targeting the seed region of oncomiRs-10b and -21. sγPNAs were encapsulated in bioadhesive PLA-HPG-CHO NPs (bioadhesive NPs; BNPs) and were delivered via CED in orthotopic GBM mouse models (Wang et al. 2022). We found that suppression of oncomiRs-10b and -21 resulted in the down-regulation of genes associated with PI3-Akt, HIF-1, and focal adhesion pathways, which sensitized the glioma cells toward TMZ-induced apoptosis. Further, a single treatment of U87- and G22- (patient-derived xenograft) based orthotopic mouse models with the combination of BNPs and TMZ resulted in tumor-free survival of animals beyond 120 d. We also evaluated the off-target effects of antiseed PNAs by intersecting mRNA sequences containing the PNA cognate sites with differentially expressed genes (DEGs) obtained from RNA sequencing analysis. We found <0.5% DEGs with the potential off-target activity of short antiseed PNAs.

PNAs have also been used to inhibit miRNAs overexpressed in kidney fibrosis, cardiovascular, and hepatic disorders. miR-33, a major regulator of lipid metabolism, plays a role in the development of kidney fibrosis via regulating fatty acid oxidation and in atherosclerosis through modulation of cholesterol flux in macrophages. Importantly, both kidneys and atherosclerotic plaques have acidic microenvironments which have been exploited to achieve targeted delivery of anti-miR-33 PNA using pHLIP. PNA-mediated inhibition of miR-33 in vivo resulted in the up-regulation of its downstream targets involved in fatty acid oxidation with reduced lipid accumulation and progression of kidney fibrosis (Price et al. 2019). Similar results were reported in an atherosclerotic mouse model where preferential delivery of pHLIP antimiR-33 PNA to the macrophages showed increased collagen and decreased lipid accumulation, resulting in atherosclerotic regression (Zhang et al. 2022).

miR-141-3p has been shown to increase after ischemic stroke and post-stroke isolation. Anti-miR-141-3p PNAs delivered via PLGA nanoparticles inhibited miR-141-3p and had neuroprotective effects in a stroke mouse model (Dhuri et al. 2021b). Elevated levels of miR-122 have been observed in patients with diabetes and obesity. Systemic delivery of miniPEG γPNA targeting miR-122 in mice fed a high-fat diet (HFD) resulted in reduced blood glucose levels and rescued the mice from endothelial dysfunction (Gaddam et al. 2022). Multiple miRNAs, that is, miR-145-5p, miR-101-3p, and miR-335-5p, have been reported to regulate the expression of the CFTR gene which encodes a chloride channel transporter. Simultaneous inhibition of miR-145 and miR-101 using antimiR PNAs showed maximum up-regulation of CFTR protein in Cacu-3 cells (Papi et al. 2022).

PNAs for infectious diseases

Since their discovery, PNAs have been extensively explored as anti-infective agents against viral and bacterial genomes. PNAs have been designed to target viral RNA templates to block the translation of reverse transcriptase which is critical for viral replication. PNAs targeting the long terminal repeats of the Hepatitis B virus have shown efficacy in cell culture as well as Hepatitis B mouse models (Zeng et al. 2016; Ghaffari et al. 2019).

PNAs have also shown promising results in the identification of blood-borne bacterial and fungal pathogens with >95% sensitivity (Nolling et al. 2016). The high sensitivity of PNAs in diagnostic applications for bacteria has warranted their use as an antibacterial agent. The major hurdle of using PNA as an antibacterial is the limited permeability of PNA across the lipopolysaccharide-rich bacterial membranes. Hence, CPP conjugated PNAs have been used to improve bacterial transfection. Antimicrobial PNA–CPP conjugates have been developed to target mRNAs involved in metabolic and biosynthesis pathways in pathogenic bacteria (Good et al. 2001). The acpP gene plays an important role in fatty acid biosynthesis and is conserved among gram-negative bacteria (Castillo et al. 2018). PNAs targeting the Shine-Dalgarno sequence and translation start site of acpP showed antimicrobial activity against Brucella suis, Haemophilus influenza, and Pseudomonas aeruginosa (Ghosal and Nielsen 2012; Rajasekaran et al. 2013; Otsuka et al. 2017). Other studies in uropathogenic E. coli have used PNA–CPP to target genes like acp, dnaB, ftsZ, and rpsH and have shown growth inhibition (Popella et al. 2022).

To improve bacterial delivery, PNAs have been conjugated with diaminobutanoic acid dendrons to target the acpP gene, which exhibited bactericidal activity against E. coli and Klebsiella pnuemoniae at 0.5 µM concentration (Iubatti et al. 2022). Another strategy to improve the delivery of PNA is conjugation with vitamin B₁₂ as it is an essential nutrient for bacteria (Wierzba et al. 2021). Improvement of antimicrobial activity was accessed by a decrease in the minimum inhibitory concentration (MIC) which is the lowest concentration of the antibiotic needed to inhibit bacterial growth. PNAs targeting folA and folP, essential for folate biosynthesis, inhibited the growth of Escherichia coli. The combination of antimicrobial PNAs with antibiotics like polymyxin B and trimethoprim showed synergistic activity and reduced the MIC of PNAs in E. coli (Dryselius et al. 2005). Synergistic action of PNAs and antibiotics was also observed when they targeted the same pathway as observed with the combination of sulfonamides and anti-folA PNA, both targeting folate biosynthesis. Antimicrobial PNAs designed to target species-specific sites demonstrated species selectivity in a mixed culture of E. coli, Bacillus subtilis, Klebsiella pnuemoniae, and Salmonella typhimurium (Mondhe et al. 2014). Anti inhA PNAs in combination with permeabilizing drugs like ethambutol showed a 1.3 log-fold clearance against intracellular M.tuberculosis (Cotta et al. 2022). Hence, combining PNAs with antibiotics is a promising strategy for treating multidrug-resistant bacterial infections.

Potential for PNA improvements in clinical applications

siRNAs were discovered in 1998 and it took almost two decades to receive clinical approval of the first siRNA-based antisense drug, patisiran, in 2018. The second (givosiran) and third (lumasiran) siRNA-based drugs were approved in 2019 and 2020, respectively. This was followed by the approval of inclisiran and vutrisiran in 2021 and 2022, respectively. Currently, there are seven siRNA-based drugs in phase III clinical trials (Hu et al. 2020). In contrast, PNAs, first reported in 1991, have progressed more slowly on the clinical front. The design of PNAs was guided by a computer model that replaced the sugar-phosphate backbone with a neutral amide linkage which is resistant to enzymatic degradation. The neutral backbone results in highly stable complexes of PNA with the target RNA. However, it also results in increased hydrophobicity, reduced solubility, and poor cellular permeability (Nielsen et al. 1991). Chemical modifications at the gamma position of the PNA backbone have helped resolve some of these limitations. Since the initial report of γPNA, it took almost a decade to establish their antisense activity in vitro. γPNAs are hydrophilic with higher solubility and exhibit superior binding affinity toward the target RNA in comparison to the unmodified PNA (Dragulescu-Andrasi et al. 2006). Further improvements in PNA design, like the tail-clamp model, contribute toward increasing the potency of antisense γPNAs (Dhuri et al. 2021a). The combination of design innovation and chemical-modified PNAs have demonstrated the therapeutic potential of PNAs to efficiently target coding and noncoding RNAs for a myriad of malignant and nonmalignant disorders (Fig. 3). Several preclinical studies have now established the efficacy and safety of chemically modified antisense PNAs in mouse models for cancer therapy (Sahraei et al. 2019; Kaplan et al. 2020), kidney fibrosis (Price et al. 2019), and atherosclerosis (Zhang et al. 2022). However, additional studies in nonhuman primates are needed to warrant the successful translation of antisense PNAs into the clinic.

FIGURE 3.

Therapeutic applications of peptide nucleic acids (PNA). Antisense PNAs can be used for targeting mRNA, pre-mRNA, and miRNA.

Among the diverse applications of antisense PNAs, infectious diseases represent a promising and upcoming therapeutic area. PNAs are effective as antibacterial agents in vitro, but they still have relatively high MIC. Hence, γPNAs can be used to lower the MIC and improve efficacy. Another advantage of PNAs as antibacterial agents is that they are not substrates for efflux pumps and resistance would not develop in that manner. γPNAs are excellent candidates for next-generation antimicrobial agents which can be further combined with current therapies to achieve synergistic activity.

Perspective on the advantages of PNA over MOE, PMO, and LNA

Multiple splice-switching PMOs have been approved by the FDA for the treatment of DMD. PMOs are administered systemically at high doses with limited toxicity. Similar to PMOs, PNAs also have a neutral backbone and can be used as therapeutic agents to induce exon skipping. In vivo evaluation of PNAs and PMOs targeting exon 23 of dystrophin protein in a mouse model indicated comparable efficacy (Yin et al. 2008). Unlike PMOs, PNAs have better synthetic accessibility to conjugate with ligands or peptides via disulfide or thioether linkages. Hence, PNA conjugates can be easily designed to achieve targeted delivery to help expand pre-mRNA targeting to other disorders.

LNA-based ASOs have advanced clinically, but reports of hepatotoxicity have posed a major setback (Burel et al. 2016). The specific composition and LNA nucleotide position are being evaluated to reduce the hepatotoxic potential (Hagedorn et al. 2018). PS LNA ASOs have been associated with thrombocytopenia, nephrotoxicity, and hepatotoxicity (Janssen et al. 2013). Only PS-modified ASOs activate an innate immune response (Frazier 2015) and induce site injection reactions post subcutaneous administration (Crooke 2007; Mansoor and Melendez 2008). Immune activation, in part, is caused by the presence of the PS moiety which increases the lipophilicity, resulting in enhanced interaction with the immune receptors and intracellular proteins to trigger cellular toxicities (Shen et al. 2019; Crooke et al. 2020). Introduction of 2′O-MOE modification has been reported to reduce the proinflammatory immune response of PS ASOs while improving their potency and pharmacokinetic properties (Monia 1997). A recent study reported that PS ASOs form a complex with toll-like receptor 9 (TLR9) which is endocytosed to activate the innate immune response (Pollak et al. 2022). However, the neutral backbone of PNAs poses minimal risk of immune activation (Upadhyay et al. 2008). Further, to our knowledge there are no reports of PNA triggered immune activation in vitro or in vivo. The preclinical efficacy and safety of chemically modified PNAs have been demonstrated against multiple targets in various mouse models. Moreover, the ability of PNAs to evade immune surveillance imparts a significant advantage over existing ASOs. Overall, the advancement of delivery platforms and chemical modifications has brought PNAs to the forefront of clinical development. Significant efforts are ongoing to realize the clinical translation of PNAs as antisense drugs, and in the next few years we expect to see safety and efficacy studies of antisense PNAs in nonhuman primates and in humans.