Non-Coding RNAs and Innovative Therapeutic Strategies to Target the 5’UTR of SARS-CoV-2

After the increasing number of SARS-CoV-2 infections all over the world, researchers and clinicians are struggling to find a vaccine or innovative therapeutic strategies to treat this viral infection. The SARS-CoV infection that occurred in 2002, MERS and other more common infectious diseases such as HCV, led to the discovery of many RNA-based drugs. Among them, siRNAs and antisense LNAs have been demonstrated to have effective antiviral effects both in animal models and humans. Owing to the high genomic homology of SARS-CoV-2 and SARS-CoV (80-82%) the use of these molecules could be employed successfully also to target this emerging coronavirus. Trying to translate this approach to treat COVID-19, we analyzed the common structural features of viral 5’UTR regions that can be targeted by non -coding RNAs and we also identified miRNAs binding sites suitable for designing RNA-based drugs to be employed successfully against SARS-CoV-2.


Introduction
After the first reports of a novel severe acute respiratory syndrome coronavirus (SARS-CoV-2) in Wuhan city (Hubei, China) the World Health Organization declared the COVID-19 outbreak a Public Health Emergency of International Concern and more recently a pandemic, as the number of cases all over the world has dramatically increased. As with other acute respiratory illnesses, the clinical presentation may vary including influenza-like symptoms to severe pneumonia or breathing difficulties, and sometimes death, more frequent in older people or individuals with pre-existing diabetes and heart disease conditions (WHO, https://www.who.int/ith/diseases/sars/en/). Similarly to the SARS-CoV outbreak in 2002, the scientific community is struggling to find novel therapies, as neither vaccines or specific treatments are currently available. Among the first therapeutic options that have been employed since the beginning we recall the antiviral therapies [1]. Several antiviral drugs such as Ribavirin (routinely employed to treat hepatitis C) [2], Nelfinavir or a combination of Lopinavir and Ritonavir (widely used as HIV protease inhibitors) [3] and Remdesivir (a monophosphoramidate nucleoside) [4] alone or in combination with chloroquine/hydroxychloroquine [5] are currently being assessed for the treatment of COVID-19. More recently, combinations of drugs (i.e., Bromhexine/Arbidol/Interferon α2b, Hydroxychloroquine/Azithromycin, etc.), biological drugs (i.e., convalescent plasma, MSC and MSC-derived exsosomes, CAR-NK cells, mRNA-1273, etc.), chinese medicine drugs (i.e., YinHu QingWen decoction, Xiyanping injection) and supplementations with Vitamins C and D are currently under investigation (https://www.clinicaltrials.gov/). However, some results about the lopinavirritonavir treatment confirmed that the combined drugs are ineffective [6].
Among all the genomic regions of the SARS-CoV genome that have been studied so far, we think that the 5'UTR region and specific portion of it should be considered crucial when devising novel therapeutic molecules also for SARS-CoV-2 infections.

5'UTR and the leader sequence of SARS-CoV
In the years following the outbreak of SARS-CoV, genomics, phylogeny, antigenic structure, immune response and potential therapeutic interventions have been reviewed [10]. Specific coding regions of the viral genome that encoded proteins fundamental for virus replication and transmission (i.e., replicases, S and M glycoproteins, envelope E protein, etc.) have been investigated as well. Some of these studies focused also on the structure and functional role of the 5'UTR region in determining coronavirus infection and replication. The 5'UTR (as well as the 3'UTR) genomic sequence is crucial for viral RNA replication and transcription [11]. In particular, a specific portion of 50-100 nucleotides at the 5' end of the genome, referred as the 'leader' sequence, has been found also at the 5' ends of all encoded transcripts (i.e., subgenomic mRNAs). This leader sequence and a cis-acting element termed transcriptionregulatory sequence (TRS) immediately following the leader sequence, represent a unique feature of coronaviruses and some other viruses of the order Nidovirales [12].
Although the viral transcription mechanism is not fully understood, two major models have been proposed: a leader-primed transcription [13,14] or a discontinuous transcription during minus-strand synthesis [15,16]. Without going into details on these mechanisms, we outline the importance of the 5'UTR region and of the highly conserved leader sequence of coronaviruses especially when prompted to find therapeutic solutions for SARS-CoV-2.
In fact, one of the most interesting aspects reported by Zeng and collaborators in 2003 was the identification of a leader sequence of 63 bp ( Figure 1A) and an intergenic sequence (IGS) of 9 bp (5'-TAAACGAAC-3') in the 5'UTR region of SARS-CoV HK-39 (NCBI Accession AY278491) [17]. Similarly, Li and collaborators reported a leader sequence of 72 bp for SARS-CoV BJ01 (NCBI Accession AY278488) coincident to that of SARS-CoV HK-39 ( Figure 1A). [18].
Also in SARS-associated coronaviruses, such as the Tor2 isolate, the presence of a core sequence (5'-CTAAAC-3') within the IGS region of the 5'UTR has been confirmed [19]. These regions have been indifferently termed IGS or TRS in different works, but the genomic location is the same.
Therefore, all of these data, while emphasizing the conservation of this genomic region among various coronavirus isolates, are useful to address the research toward the development of effective strategies aimed at reducing or abolishing viral activity and replication in vitro and in vivo. In fact, many strategies that we will describe in the following paragraphs have been adopted to target these viral genomic regions.

RNAi against the 5'UTR and the leader region of SARS-CoV
RNA interference (RNAi) was one of the strategies adopted in the past to inhibit the replication of coronavirus [20,21]. Several in vitro studies demonstrated that siRNAs targeting different region of SARS-CoV genome were effective in reducing the expression of viral sub-genomic mRNAs, ultimately leading to an inhibition of viral replication. siRNA directed against the Spike region was the most effective although the authors demonstrated that also those siRNAs targeting the leader sequence or the TRS region and the 3'UTR were able to prevent SARS-CoV infection in Vero-E6 cells ( Figure 1A) [20]. The SARS-CoV Spike protein was considered by many authors the most obvious candidate owing to its fundamental role in infecting human respiratory epithelial cells by interacting with human receptors (i.e., ACE2) [22].
Similarly, other studies reported the effective targeting of different regions such as the SARS-CoV M gene [23].
However, Li et al. by employing a specific siRNA (outlined in Figure 1A) were able to target the SARS-CoV leader sequence and effectively inhibit virus replication in Vero E6 cells [24].
However, a detailed study demonstrated that specific siRNA duplexes (i.e., siSC2 and siSC5), targeting the SARS-CoV Spike and ORF1b (NSP12) regions can significantly suppress SARS-like symptoms also in vivo (i.e., in infected macaques) [24]. More importantly, the RNAi approach displayed no adverse effects and was demonstrated to be effective for both prophylaxis and therapy [25].
Although this approach cannot be considered an antiviral approach (as animals developed symptoms), it is surely a valid strategy to reduce the viral load and the severity of the disease. In any case, the idea that RNAi can be used to mitigate the symptoms of new coronavirus infections and other emerging infectious diseases was a debated concept in the past [26], but it is still actual and valuable of further investigations. In fact, we believe that approaches based on the common features shared between SARS-CoV and SARS-CoV-2 could be employed to block the virus replication.

5'UTR and the leader sequence of SARS-CoV-2
To evaluate whether the concepts already found for 5'UTR of SARS-CoV can be transferred also to SARS-CoV-2, we calculated the similarity between the 5'UTR of the two coronaviruses. We found that the similarity of these two regions is 88.76% ( Figure 1B) [27]. This sequence is very similar to that of SARS-CoV ( Figure 1B) and this strongly suggests that the RNAi strategy already adopted for SARS-CoV and focused on this specific region would likely lead to inhibition of SARS-CoV-2 replication. However, RNAi is not the only strategy that can be employed to this purpose.

Deletion of specific regions in the 5'UTR of SARS-CoV
In 2006 to test the promoter activity of the 5'UTR of SARS-CoV in eukaryotic cells and identify the fundamental regions necessary for viral replication, Zhang and collaborators assessed the function of many 5'UTR clones lacking increasing portions of the SARS-CoV 5'UTR sequence [28]. They found that the wild type SARS-CoV 5'UTR has a promoter activity in eukaryotic cells such as A549, HepG2, ECV304, HeLa and Vero E6 and that deletion mutant plasmids with different 5'UTR length have different activities. Interestingly, the mutant plasmid lacking the region 1-36 did not altered the activity, whereas the plasmid lacking almost all the UTR (i.e., the 1-222 region) abolished completely the SARS-CoV promoter activity in human cells. In fact, it is long been known that subgenomic mRNAs lacking the 5' leader sequence are not able to replicate [29,30]. At least four stem loop structures are located in this 5'-end region of the coronavirus genome and these secondary structures are actively implicated in viral replication and transcription [31]. Moreover, the authors identified the initial site of transcription at the 56 th nucleotide (see Figure 1A) that is proximal to the TRS sequence and to those regions already targeted by siRNAs. It is reasonable to think that this position is among the crucial genomic regions to consider for targeting purposes.

Secondary structures of coronaviruses 5'UTR
The RNA synthesis processes of coronaviruses takes place in the cytoplasm and is regulated by proteins of the host cell. Among these proteins, the zinc-finger and RNA binding motif MADP1 was able to interact with the 5'UTR region of the SARS-CoV genome, and this interaction was confirmed also in coronavirus infectious bronchitis virus (IBV) [32]. Interestingly, the interaction seems The presence of these stem-loops is highly conserved in many coronaviruses derived from all three major CoV groups [33]. Moreover, a couple of papers pointed to the importance of single-nucleotide mutations and deletions that are able to destabilize the first two stem loops in the 5'UTR of coronaviruses, thus inhibiting viral replication [34,35]. By analyzing the first 100-nucleotide region of the 5'UTR of IBV, MERS, SARS-CoV and SARS-CoV-2, we confirmed their similarity (Figure 2) and we hypothesized that also their function could be almost identical (i.e., enhancement of viral replication) [11]. However, this similarity extends also to the entire 5'UTR of SARS-CoV and SARS-CoV-2 ( Figure 3).
Therefore, all of these data outline once more the importance of 5'UTR when devising therapeutic strategies for coronaviruses.

Locked nucleic acid antisense oligonucleotides (LNA)
RNAi is not the only strategy that can be adopted to target the 5'UTR of coronaviruses. A few years ago an innovative antiviral strategy for HCV, another common widespread virus, based on a locked nucleic acid-modified DNA phosphorothioate antisense oligonucleotide, named Miravirsen, has been proposed [36]. This approach was based on a high affinity antisense oligonucleotide complementary to the human mature miR-122, a liver-specific and highly expressed microRNA. This microRNA is able to bind the 5'UTR of the HCV genome in two distinct regions and to promote actively the propagation of HCV RNA [37]. Prolonged subcutaneous administration of To translate this approach into the identification of therapeutic molecules for the potential treatment of COVID-19 infections, we searched for potential miRNAs binding sites targeting the 5'UTR of SARS-CoV-2. Target recognition algorithms such as PITA [38] allowed identifying the accessible genomic binding sites for human miRNAs. Many miRNA binding sites were identified in conserved regions (Figure 4) that were filtered according to their score (ddG Similarly, we focused to find highly expressed miRNAs in the lungs that are the main target organs for SARS-CoV-2. Therefore, we extracted and analyzed the expression data contained in the human miRNA tissue atlas [39] to gain information about the quantity of these miRNAs in human tissues. We found that miR-4507 was among the most expressed miRNAs (99 th percentile) in the lungs together with miR-638 (97 th percentile) which is highly expressed also in the pleura (99 th percentile). Two distinct binding regions were found in the 5'UTR of SARS-CoV-2 genome for miR-3150b-3p and miR-602. We still do not know if these miRNAs have crucial functions in promoting viral replication or if the binding of other miRNAs in different 5'UTR region has inhibitory functions. If we suppose a similar enhancing mechanism, we should hypothesize that the design of antisense oligonucleotides similar to Miravirsen could be able to sequester these miRNAs and inhibit viral replication. In any case, these findings prompt the research to validate these targets, explore their function and potentially identify novel drugs.
However, many other miRNAs resulted to bind coding regions and 3'UTR of the coronavirus (see Supplementary Material) where other cis-acting region are present, although the regulatory mechanisms underlying these interactions are still not completely understood [40]. The RNA-based antisense therapy is of wide applicability and the strategy herein suggested was already assessed in chronic HCV infections with no major adverse events. Based on these findings, we recommend to consider the possibility to employ miRNAs and their antisense analogues as effective candidates for targeting coronaviruses and to identify novel molecules for antiviral therapy.

LNA-based oligonucleotides for in vivo applications (GapmeRs)
In the last few years, novel single-stranded antisense oligonucleotides, named  Figure 5B).
Therefore, these data emphasize that GapmeRs could be optimized (i.e., proper binding region, sequence length, prevention of off-target effects) and effectively employed to inhibit the replication activity of SARS-CoV-2.

Conclusions
We have summarized and discussed here the panorama of non-coding RNA molecules that can be employed to target the 5'UTR region of SARS-CoV-2 with the aim to find novel drugs or innovative therapeutic strategies. The process to commercialize a novel drug is lengthy and costly owing to the regulatory processes related to pre-clinical and clinical studies. Despite this, we think that many of the molecules already studied and developed after the previous SARS- Independently by the use of siRNAs, miRNAs or GapmeRs, the arsenal that can be employed to fight SARS-CoV-2 is promising and potentially very powerful. In principle, some of these therapeutic options could alleviate COVID-19 symptoms or reduce viral replication. We strongly believe that none of these possibilities should be left unassessed, as the discovery of an effective vaccine could be a very long and uncertain path.