Noncoding RNAs in cancer therapy resistance and targeted drug development

Cancer is an unconquered disease that generally causes mortality and morbidity worldwide and generates many adverse socioeconomic effects. Although protein-relevant therapeutics such as antibodies against Programmed Cell Death 1 (PD1), programmed death-ligand 1 (PDL1), and cytotoxic T-lymphocyte-associated protein 4 (CTLA-4) have driven a revolutionary trend in pharmacotherapy and drug development, some protein targets encoded by oncogenes are undruggable or inadequate for achieving remission, and cancer cells can acquire drug resistance [1]. Therefore, the treatment of cancer further requires multiple types of targets involved in oncogenic pathways for successful intervention.

To advance the understanding of cancer initiation and progression, many genomic and proteomic approaches have been developed [1‐3]. The landscape of genomic mutations in cancer reveals that many mutations or copy number changes in cancer are frequently located in noncoding DNA regions [1, 4]. Noncoding DNA has been reported to cover 95% of DNA sequences in the human genome, most of which are transcribed into tens of thousands of functional noncoding RNAs (ncRNAs), including microRNAs (miRNAs), small interfering RNAs (siRNAs), antisense RNAs (asRNAs), and long noncoding RNAs (lncRNAs) [4‐7]. Recent studies have also reported a novel type of ncRNA, circular RNA (circRNA) [8‐10]. A large portion of circRNAs are generated from exons of coding genes, and most do not express protein [8‐10]. The biogenesis of several kinds of ncRNAs, such as miRNA [1, 2, 5], short hairpin RNA (shRNA) and siRNA [11], lncRNA [4], and circRNA [8‐10], is summarized in Fig. 1. Accumulating evidence shows that ncRNAs are dysregulated and implicated in various cancer processes, such as cancer stem cell (CSC) initiation, metastasis, and drug resistance, highlighting the role of ncRNAs as potential therapeutic targets in cancer [5, 6, 8‐15]. Several miRNAs have reached clinical trials [15‐17]. In addition, lncRNAs and circRNAs have demonstrated significant clinical relevance in cancers due to their relatively complex and diverse structures and functions acting through multiple mechanisms [6, 8]. Moreover, the preclinical studies and increased success rates of nucleic acid therapeutics provide an opportunity to target ncRNAs for cancer treatment [5, 15‐17].

Here, we summarize ncRNAs in therapeutic resistance, the potential as therapeutic targets, the current status of ongoing clinical trials, and therapeutic approaches for targeting ncRNAs. We also discuss the challenges for the efficient delivery of ncRNAs as therapeutics, the obstacles in clinical trials, and the perspective for the future design of nucleic acid therapeutics.

In addition to their roles in chemoresistance, ncRNAs play a part in radiosensitivity as summarized in Fig. 2b. As tumor cells generally impair DNA damage repair ability and have the propensity for more rapid division, these cells are more prone to apoptosis induced by radiation than normal tissues [42]. However, several findings have implied that radiotherapy might concomitantly promote the recurrence and metastasis of malignant tumors by activating epithelial–mesenchymal transition (EMT) and/or generating CSCs [43, 44]; importantly, ncRNAs are closely involved in these processes.

Several studies have revealed the strong relation between the expression patterns of a subgroup of miRNAs with the radiotherapy response in various cancers, including non-small cell lung cancer, head and neck tumors, squamous cell carcinoma, and squamous cervical carcinoma [45‐48]. Thus, these miRNAs could be predictors of radioresistance (Fig. 2b, left panel). For example, Marina et al. found that several target genes of miR-139-5p were strongly predictive of outcome in radiotherapy-treated patients, suggesting that this miRNA may be a potentially useful predictive biomarker of radioresistance in breast cancer [46]. Besides, some miRNA expression levels change in a unique pattern during radiotherapy treatment. Thus, miRNAs can also be used as monitors to evaluate the real-time response to radiotherapy and to prevent delays in changing to an alternative treatment regimen [49‐51]. The reported “circulating miRNAs” and plasma lncRNAs or lncRNAs such as HOTAIR, metastasis-associated lung adenocarcinoma transcript 1 (MALAT1), H19, and GAS5 might also have value in the evaluation of radioresistance in cancer patients [52‐55].

Mechanically, ncRNAs acting as modulators of radiosensitivity are mainly through regulating the processes of apoptosis, DNA damage repair, and EMT [52, 56]. For example, miR-21, miR-125b, miR-181a, miR-196a, and miR-148b suppress the expression of the apoptosis-related genes caspase-3, intercellular adhesion molecule-2 (ICAM-2), Protein Kinase C Delta (PRKCD), annexin A1 (ANXA1), or DNA methyltransferase 3b (DNMT3B) in a wide spectrum of cancers [56]. LncRNAs, such as LincRNA-p21, LOC285194, ANRIL, AK294004, LncRNA-ROR, and MALAT1, can regulate apoptosis-related genes either by binding to the protein partners or by acting as competing endogenous RNAs (ceRNAs) [57‐59] (Fig. 2b, right panel). Some lncRNAs could also modulate radioresistance by regulating the DNA damage response [56, 60, 61]. For instance, lncRNA LINP1 can enhance repair of DNA double-strand breaks by serving as a scaffold linking Ku80 and DNA-dependent protein kinase catalytic subunit (DNA-PKcs). Blocking LINP1 increases the sensitivity of the tumor-cell response to radiotherapy in breast cancer [61]. Furthermore, many EMT- and CSC-related lncRNAs are also responsible for radioresistance [62‐64]. MALAT1 is the one reported to promote radioresistance through triggering EMT and CSC activity [63, 64]. Silencing the expression of MALAT1 sensitizes nasopharyngeal carcinoma cells to radiotherapy both in vitro and in vivo through the inactivation of EMT and CSCs by modulating the miR-1/SLUG axis [64]. Some other lncRNAs, such as LincRNA-p21, NEAT1, LncTCF7, and Lnc34a, might also affect radiotherapy responses [65‐67]. These findings highlight the effects of ncRNAs on CSCs and radioresistance, supporting the design of effective strategies to improve radiotherapy responses.

In addition to the involvement of ncRNAs in therapy resistances as discussed above, numerous ncRNAs have also shown potential as therapeutic targets because of their differential expression patterns between cancerous and normal tissues and their important roles in carcinogenesis [2, 5, 68‐70]. With advances in nanotechnology, numerous clinical trials on RNA-guided precision medicine have recently been developed and/or are ongoing [2, 16, 71]. As summarized in Table 1, among the annotated ncRNAs, miRNAs are the most extensively investigated. MiRNAs function as either oncogenes or tumor suppressors, which, in turn, cause aberrant translational inhibition or degradation of their target mRNAs [72]. The pivotal double-faced roles of miRNAs in cancer make them as both therapeutic candidates and the direct therapeutic targets [17]. MiRNA mimics and anti-miRNAs are two major therapeutic forms used to mimic the function of endogenous tumor suppressor miRNAs and to deplete oncogenic miRNAs, respectively [73, 74]. For instance, miR-34a mimic was the first miRNA-based therapy to be introduced into the clinic [16]. MiR-34a can antagonize many different oncogenic processes by silencing oncogenes, including cyclin-dependent kinase 4/6 (CDK4/6), SIRT1, and SOX2, which function in self-renewal, migratory potential, and chemoresistance in various types of cancers [16, 75, 76]. In a phase I clinical trial, researchers designed a liposomal nanoparticle to deliver corresponding miR-34a mimics to the liver and precisely target cancer tissues [77]. More excitingly, many clinical trials of miRNA-based therapeutics have reached phase 3 or 4 ( https://​clinicaltrials.​gov), suggesting the promise of these therapeutic targets in precision medicine. Screenings of miRNA-based drugs have identified specific miRNA targets in different cancers, including miR-31-3p and miR-31-5p in colorectal cancer (CRC) [78], and miR-21 and miR-200 in oral cancer [79, 80]. The ongoing clinical trials have indicated that miRNA could be a widespread target in cancer treatment.

Apart from the extensively studied miRNAs, lncRNAs and circRNAs have recently emerged as novel targets [6, 8, 9, 68]. Compared with miRNAs, lncRNAs and circRNAs act through more diverse mechanisms in carcinogenesis [4, 81, 82]; thus, targeting lncRNAs and circRNAs provides varied means to modulate a range of critical processes in cancer development. Double-stranded RNA-mediated interference (RNAi) and single-stranded antisense oligonucleotides (ASOs) are two major approaches to target lncRNAs. ASOs can reduce the levels of oncogenic isoforms of lncRNAs by regulating alternative splicing, modulating RNA–protein interactions, or causing lncRNA degradation [83, 84]. For example, targeting lncRNA MALAT1 with ASO induced differentiation and inhibited metastasis in a mouse model of breast cancer [84, 85]. The antimetastatic effect of MALAT1 targeting by ASO was also reported in a lung cancer xenograft model, highlighting the potential of MALAT1 as a therapeutic target in multiple tumors [86]. Notably, a subset of lncRNAs named natural antisense RNAs (NATs) are sometimes located near important tumor suppressors, such as ANRIL and p21-AS [87, 88]. Therapeutic inhibition of cis-acting NATs with a special type of ASO, antagoNATs, can potentially upregulate the overlapping tumor suppressor genes; thus, modulating lncRNA expression could be a tool to regulate gene expression. However, therapeutic targeting of lncRNAs and circRNAs remains mainly at the laboratory stage.

Approaches for therapeutic targeting are essential for precision medicine. Several preclinical studies have been initiated to investigate anticancer strategies for targeting oncogenic ncRNAs (https://​clinicaltrials.​gov). Three strategies have been proposed: ASOs, locked nucleic acids (LNAs), and morpholinos [1, 89‐94]. ASOs are single-stranded oligonucleotides that have specific complementarity to target sequences to promote target RNA degradation by RNase H as shown in Fig. 3(a) [91]. LNAs are also single-stranded oligonucleotides containing a stretch of DNA flanked by LNA nucleotides and offer specific complementarity and RNase H-mediated degradation of the target sequence as shown in Fig. 3(b) [92, 93]. Different from ASOs and LNAs, morpholino oligonucleotides (MO) are 25-nt nonionic DNA analogs used to promote RNA degradation through binding target RNAs in diverse organisms (Fig. 3(c)) [94‐97]. These strategies have been applied for targeting oncogenic ncRNAs in cancer. For instances, miR-10b ASOs together with a low dose of doxorubicin showed a significant decrease in tumor size compared with the results only using doxorubicin monotherapy to treat breast cancer in mouse models [98]. The researchers also applied miR-10b LNAs for investigation and found that miR-10b LNAs can enhance the sensitivity of breast cancer to doxorubicin in mouse models, with no additional damage to normal tissue, suggesting low toxicity associated with the delivery of this LNA nanoparticle [98]. MALAT1 ASOs could also inhibit the metastasis of cancer cells and the tumor burden in mice [99]. AVI-4126, a morpholinos-based drug, was used to inhibit c-MYC translation in a sequence-specific manner by simultaneously blocking the expressions of c-MYC and causing the mis-splicing of its pre-mRNA, resulting in significant growth inhibition in various cancer cells, such as prostate cancer, breast cancer, and lung cancer [95, 97, 100]. Chang et al. designed a phosphorodiamidate morpholino oligomer that effectively silenced miR-487a in a mouse model and reduced tumor growth and metastasis [96]. These studies indicate that morpholinos-based drugs of targeting oncogenic ncRNAs may represent a promising approach for cancer therapy. Further clinical trials are required.

With unique properties in their chemical behavior, modes of action and clinical pharmacokinetics, ncRNAs have been considered not only as the therapeutic targets, but also as an alternative promising therapeutic tool for cancer treatment. This possibility is especially meaningful for small ncRNAs, including miRNAs (Fig. 3(d)) [2], siRNAs, shRNA (Fig. 3(e, f)) [101], small activating RNAs (saRNAs) [7], guide RNAs (gRNAs) (Fig. 3(g)) [102], and other forms of small RNAs (sRNAs) [1, 6].

SiRNA is a well-studied small ncRNA tested for preclinical trials. SiRNAs are a class of synthetic short double-stranded RNAs with 21 base pairs that are recruited by the RNA-induced silencing complex (RISC) in cells and silence the expression of their target mRNAs, lncRNAs, or circRNAs by complementary base pairing (Fig. 3(e)) [101, 103]. Zorde et al. developed a system for local prolonged effects of siRNA (Local Drug EluteR, LODER) enabling the utilization of siRNAs against mutated Kirsten rat sarcoma (KRAS) (siG12D LODER) [104]. Mutated KRAS has been reported to be resistant to small molecule drugs that target oncogenic proteins. They found that siG12D LODER suppressed the growth of human pancreatic cancer cells in vivo and therefore prolonged mouse survival [104]. In an open-label phase 1/2a study, an improved therapeutic effect was reported in patients with locally advanced pancreatic cancer (LAPC) treated with a combination regimen of siG12D-LODER and gemcitabine [105]. Some other siRNA-based drugs have already completed phase II clinical trials as shown in Table 2, including DCR-MYC for MYC knocking down to arrest cell growth in hepatocellular carcinoma (HCC) (NCT02314052), Atu027 for PKN3 knocking down to regulate cell migration in metastatic pancreatic adenocarcinoma (NCT01808638), etc. The rapid development of siRNA- and miRNA-based clinical trials profits from the shortness of small RNA sequences and technical advances in previously introduced delivery carriers. These advantages facilitate enduring and safe circulation in the blood and ease of uptake by target cells to improve the biodistribution and bioavailability of these RNAs during trafficking to cancer cells.

saRNA is another type of small double-stranded ncRNA designed to target gene promoters to activate transcription and thus upregulate gene expression [7]. In a preclinical trial, Reebye et al. designed a saRNA targeting CEBPA that can activate the transcription of C/EBP-α in a liver cancer model [106]. They found that intravenous injection of C/EBPα-saRNA reduced the tumor burden and suppressed the expression of interleukin (IL) 6R and c-Myc and inhibited STAT3 phosphorylation [106]. The delivery efficiency was improved by loading C/EBPα-saRNA in a liposomal nanoparticle in the following clinical trial study [107]. The ongoing clinical trial indicated that saRNAs are promising for activating the transcription of tumor suppressor.

gRNA is also an important type of small ncRNAs. Since the discovery of the clustered regularly interspaced short palindromic repeats (CRISPR)–CRISPR-associated (Cas) system, Cas-gRNA-based biotechnology has developed rapidly and massively (Fig. 3(g)) [102, 108]. Diverse RNA-programmable CRISPR–Cas enzymes have been gradually found and subsequently applied to reverse the aberrant expression of oncogenes and tumor suppressor genes [102]. Generally, the CRISPR–Cas system has two advantages over other gene editing strategies. First, the CRISPR–Cas system offers sequence-specific DNA targeting through a single-guide RNA (sgRNA)-based nucleoprotein complex that specifically cleaves the genomic DNA of interest to accomplish gene editing and mutation. Second, designing the sgRNA sequence targeting the desired DNA sequence is simple and flexible. Due to the great advantages of the sgRNA-guided CRISPR–Cas system, CRISPR systems have been widely adapted to facilitate the discovery of new targets in cancer therapy. For example, Yamauchi et al. performed a genome-wide CRISPR–Cas9 screen in AML cell lines to identify novel targets for AML therapy and discovered that the mRNA decapping enzyme scavenger (DCPS) gene, which is involved in pre-mRNA metabolic pathways, is essential for AML cell survival [109]. More interestingly, these researchers further found that germline biallelic DCPS loss-of-function mutations resulted in failure to induce leukemogenesis in humans, suggesting that CRISPR–Cas9-mediated silencing of DCPS is a potential strategy for AML therapy [109]. In addition, recent applications of the CRISPR–Cas system in chimeric antigen receptor (CAR) T cells, including CD133-specific CAR T cells with PD-1 deficiency and CD3-specific CAR T cells with diacylglycerol kinase (DGK) deficiency [110, 111], have been shown to be promising strategies in cancer immunotherapy. Although most studies exploring the CRISPR–Cas system are still in the preclinical stage, multiple Cas-based clinical trials are in progress or will be commencing soon.

Although small ncRNAs have been shown to be promising and effective therapeutic drugs in vitro, the low bioavailability of these nucleic acid drugs in vivo is a major challenge [6, 73, 94]. Thus, the development of advanced drug delivery strategies is urgently needed. To overcome the general problems of a short half-life, off-target effects and low transfection efficiency in RNA delivery, various small ncRNA carriers or systems have been proposed and extensively investigated, including nanoparticles (Fig. 4a), ncRNA modification (Fig. 4b, c), and oncolytic adenovirus strategy (Fig. 4d).

The first class of carriers is nanoparticle-based and includes self-assembled oligonucleotide nanoparticles [112], lipid-based nanoparticles (LNPs) [113], inorganic nanoparticles, and polymeric nanoparticles [101]. Huang et al. reported the assembly of a calcium–phosphate–lipid nanoparticle, which showed higher efficiency in delivering siRNA into HCC cells than into normal hepatocytes [114]. These lipid nanoparticles were subsequently loaded with VEGF siRNAs and downregulated VEGF expression specifically in HCC both in vitro and in vivo, exhibiting a potent antiangiogenic effect in the tumor microenvironment in a mouse model of HCC [115]. This carrier spectrum is also reflected in the miRNA monotherapies used in several types of cancers—for example, miR-34 and miR-125b in non-small cell lung cancer [116, 117], miR-212 and miR-216b in pancreatic ductal adenocarcinoma [118, 119], miR-328 in CRC [120], and miR-221 in liver cancer [121]. Several clinical trials for siRNA-/miRNA-based therapeutics targeting cancers have been performed (Fig. 3(d, e)).

NcRNA modification is another effective strategy to improve the precision and durability of small ncRNAs in targeting genes of interest in cancer, including siRNAs chemically conjugated with carriers forming carrier-siRNA conjugates (Fig. 4b) and siRNAs modified with lipid and PEG molecules, and then self-assembled lipid nanoparticles named as SAMiRNA (Fig. 4c). In developing therapy for liver cancer, two cooperative groups have chemically conjugated special siRNAs with (2–3)N-acetylgalactosamine (GalNAc) and developed GalNac-conjugated siRNAs [122]. GalNAc is a type of ligand with high binding affinity to the asialoglycoprotein receptor (ASGPR), which is specifically expressed and localized on the surface of hepatocytes. Martin et al. demonstrated that the interaction of ASGPR with the GalNAc ligand facilitates clathrin-mediated endocytosis [123]. Finally, most GalNac-conjugated siRNAs could be precisely delivered into the lesions of HCC and form RISC complexes to degrade target mRNAs [122]. Currently, the developed GalNAc–siRNA conjugates are undergoing clinical trials. The team subsequently refined this siRNA by optimizing the positioning of the 2'-deoxy-2'-fluoro and 2'-O-methyl ribosugar modifications to enhance stability without compromising the intrinsic RNAi activity, yielding significantly improved potency and duration in preclinical species [124]. Another challenge is the cytotoxicity induced by delivery carriers such as cationic liposomes, which can trigger pulmonary inflammation and the production of reactive oxygen intermediates [125]. For instance, siRNAs conjugated to the cell-penetrating peptide (CPP) TAT(48-60) indeed exhibited improved gene silencing efficiency. However, this conjugate also concomitantly activates the innate immune response [126]. Modification of siRNAs with hypotoxicity is also a problem requiring a solution. Yoon et al. developed SAMiRNAs by conjugating siRNAs to hydrophilic polymers and lipids to form conjugates (Figs. 4c and 3 (e)) [112]. SAMiRNAs can be specifically delivered into tumor tissue with high stability and have high knockdown efficiency. The advantage of this method is that the induction of cytokines in mouse peripheral blood mononuclear cells (PBMCs) and the levels of toxicity in liver and kidneys were below the relevant detection limits [112].

The methods described above are based on transient transfection of dsRNA. However, effective siRNA targeting for therapeutic purposes would require sustaining effects of gene silencing on oncogenic RNAs of interest. The emergence of a stable RNA silencing strategy via oncolytic adenovirus has introduced a revolutionary advance in cancer therapy [127]. The oncolytic adenoviral vector is designed to remove the partial sequences of the E1A gene region to render the adenovirus replicative exclusively in cancer cells [127]. Thus, compared to conventional therapy, this engineered adenovirus showed reduced toxicity and achieved promising results (Figs. 4d and 3(f, g)) [127‐130] . The oncolytic adenoviral vector allows the insertion of large numbers of different genes with various DNA sequence lengths, including human genes, shRNA sequences, and CRISPR–Cas9 sequences. We consider Ad-shRNA as an example. Machitani et al. developed a telomerase-specific replication-competent adenovirus (TRAD), which carries the tumor-specific promoter-driven E1 gene expression cassette and exhibits increased replication efficiency and antitumor activity [130]. This group loaded shRNA targeting Dicer into the TRAD. After infection, TRAD-shDicer efficiently induced Dicer knockdown and exhibited significantly higher replication efficiency and prodeath activity in tumor cells than in normal cells [130]. Continued progress in the development of oncolytic adenovirus strategies might allow this approach to be an important and powerful alternative tool to treat cancer.

NcRNAs are emerging as crucial players in tumorigenesis. Recent progress in biotechnologies such as high-throughput sequencing, genome editing, mouse modeling, and pharmaceutical chemistry has allowed functional studies of ncRNAs to provide a new perspective for waging the war against cancer. In addition to miRNAs and lncRNAs, other novel ncRNAs, such as transfer RNA (tRNA) fragments, snoRNA-related lncRNAs (sno-lncRNAs), and circRNAs, have also begun to appear on the radar of cancer researchers. Notably, snoRNAs have regained attention in cancer research, and snoRNA derivatives might be potential players in cancer development. The tissue-specific expression of ncRNAs makes them exciting candidates for molecular targeting. Additional ncRNA targets for cancer treatment are expected to be discovered in subsequent studies. However, because ncRNAs vary in length and modes of action, the development of systematic genomic and functional approaches will be needed to better understand the roles of ncRNAs and to evaluate their potential as therapeutic targets.

As chemotherapy and radiotherapy remain the mainstream treatment approaches for cancer patients, the roles of ncRNAs in mediating chemo- and radioresistance will be increasingly appreciated. Although a subgroup of ncRNAs—to date, miRNAs and lncRNAs, in particular—have proven to be useful biomarkers for predicting treatment outcomes or monitoring therapeutic responses, most studies are still in the preclinical stage. In addition, only a few of these ncRNAs exist stably in body fluid, thus enabling a noninvasive liquid biopsy approach. More efforts are needed to discover additional circulating ncRNAs for convenient clinical diagnosis. Notably, the results of some studies evaluating the potential of ncRNAs as biomarkers are conflicting. Larger cohorts of clinical data should be mined to reconcile these controversies.

Not only are ncRNAs promising targets for treating cancer and modulating cancer treatment sensitivity, approaches for targeting ncRNAs could be RNA-based. The rapid evolution of nucleic acid therapeutics offers an exceptional opportunity to explore ncRNAs as druggable targets in the clinic. MiRNA mimics and modified miRNAs/siRNAs are currently major RNA-based drugs that target mRNAs and ncRNAs. Other methods, such as ASOs, morpholinos, and small molecules, are also promising approaches via the modulation of ncRNA degradation, alternative splicing, and RNA–protein interactions. The CRISPR–CAS system represents another prospective method, but much additional study is needed for its eventual application in the clinic. Despite the enthusiasm, several obstacles still need to be overcome. First, delivery technologies with increased efficiency should be developed; crossing the cell membrane remains the foremost issue. The complex internal microenvironment makes the delivery and application of ncRNA difficult, including RNA degradation and instability, off-target, and low transfection efficiency. Development of more advanced delivery strategy is urgency. The combination of two or more carriers may be a good choice for ncRNA targeting, such as combination of nanodesigns with organ-specific response receptor may improve the precision and efficiency of drug delivery. Second, identifying a means to evade nuclease degradation or innate immune system targeting is critical for increasing bioavailability. Last but not least, minimal off-target effects and toxicity should be ensured. Despite these challenges, nucleic acid therapeutics might be powerful drugs for cancer treatment.