1. Introduction
Since 1964, when Epstein discovered, in pediatric cases of Burkitt’s lymphoma, the first human virus linked to malignancy, the scientific attention focused on the viral regulatory mechanisms within host cell capable to create the perfect cellular environment for viral replication. Since then, it has been postulated that the 12% of all human cancers have a viral etiology [
1]. Also, novel emerging viruses, as SARS-CoV-2 and Zika, that spill over from wildlife into domestic animals and humans, are becoming a constant threat for human public health [
2]. The fact that viruses dramatically limit their genomes size to the minimum number of genes for infection [
3], suggests that viruses and their hosts have been involved in a never-ending struggle of adaptation defined by a plethora of complex mechanisms [
4]. The crucial and central point is that the viruses are perfect parasites [
5] and cannot survive without a host cell providing the transcriptional machinery. On the other side, the eukaryotic cell, does not freely allow parasites to enter, so the viruses have developed many different mechanisms in order to skip the host cell barrage, beginning with the recognition of surface molecules acting as key door, and continuing with fundamental immune evasion, all aimed at translocating proteins and genetic material into the cells. On that note, it raises that viruses are not alive stricto sensu but, even if they are made of the simplest structure existent, they result the cleverest infectious agents because of the ability to adjust themselves, use and trick the host.
The first step for a virus, either enveloped or non-enveloped, after initial attachment to the cell surface, is the penetration in the host by both membrane fusion and endocytosis [
6]. Once inside the host, a virus particle uncoats its capsid and releases the naked viral genome in order to establish gene expression and viral genome replication [
7]. It has been postulated that since the viral genome is size-limited, any non-coding space is rationated [
8]. Recently, thanks to the novel high-throughput RNA sequencing techniques it has been discovered that many, but not all viruses genome express non-coding RNAs (ncRNAs) for their own benefit [
9]. According to the Baltimore classification (first defined in 1971), viruses are divided into seven categories depending on the nucleic acid (DNA or RNA), strandedness (single-stranded or double-stranded), sense, and method of replication [
10]. The viruses that have include the majority of ncRNAs belong to a DNA virus family (herpesviruses), perhaps because they have relatively large DNA genomes [
8] capable to produce RNA intermediates [
11].
RNA viruses show a great intrinsic epidemic potential [
12] due to the high mutation rates and high frequency of recombination events. Surprisingly, it has been discovered that also RNA viruses produce ncRNAs whom expression is highly correlated with viral infection activity [
13].
The reason why the ncRNAs, and in particular viral ncRNAs, are earning more interest rely on their functions as regulators of translation, RNA splicing, and gene expression [
14] and because of the important impact they can exert on host pathways. NcRNAs belong to a very heterogeneous family, in terms of length, conformation and cellular function [
15]. At present, ncRNAs can be divided by length (cut-off 200 bp) into two main groups: long non-coding RNA (lncRNA) and small non-coding RNA (sncRNA). LncRNA can be further grouped into linear RNAs and circular RNAs.
Linear lncRNA molecules are at least 200 nucleotides long, often harbor a poly-A tail and can be spliced, as mRNAs, but lack of protein-coding potential [
16]. LncRNAs are generally nuclear localized, demonstrating their role as regulators of nuclear organization and function [
17]. Circular RNAs (circRNAs) are a novel class of non-coding RNAs, whose main characteristic is the covalently closed loop structure without terminal 5ʹ caps and 3ʹ polyadenylated tails [
18]. Since circRNAs contain miRNA response elements (MREs), it has been suggested they act as miRNA sponges through competitive binding to miRNAs, with a consequent weakening of mRNAs regulation [
19].
Small non-coding RNAs are a class of non-coding RNA composed mainly by microRNAs (miRNAs), piwi-interacting RNAs (piRNAs), small nucleolar RNAs (snoRNAs), and recently discovered tRNA-derived RNA fragments (tRFs) [
14]. MiRNAs negatively affect mRNA protein output by imperfectly base pairing with the 3′-untranslated region (3′UTR) [
20]. The complexity of this mechanism lies in the combinatorial mode of miRNA action in mRNA regulation, since a single miRNA is able to target several mRNAs [
21].
All the ncRNAs classes described above have been extensively studied in mammals, but since the first v-ncRNA was discovered, the v-ncRNAs become extremely appealing because of the role played in all the infection steps. V-ncRNAs are gaining as much pathological importance as viral structural proteins [
22]. In this review, we intent to examine how viruses have evolved a common strategy and which are the crucial host pathways targeted through v-ncRNAs in order to grant and facilitate their life cycle.
2. Viral Immune Evasion Strategies
The first step of the infection process is the evasion from the host immune surveillance system. This difficult task to achieve, is a complex balance between limiting viral gene expression in order to limit antigen presenting molecules [
23] and producing viral immune evasion against innate and adaptive immunity [
24].
Since ncRNAs are not presented via Major histocompatibility complex (MHC), they result to be non-immunogenic to the adaptive immune system and thus particularly useful tools for viruses to influence host cell functions [
25]. Interestingly, viruses produce v-ncRNAs capable of controlling not only the expression of viral genes, but also influence host cell regulation and evade host innate and specific immune responses, crucial mechanisms in the viral pathogenetic processes [
9]. Humans are equipped of pattern recognition receptors (PRRs), such as Toll-like receptors (TLRs), retinoic acid-inducible gene-I (RIG-1), and protein kinase R (PKR), that are able to detect foreign nucleic acids, as viral RNAs [
24]. Nevertheless, viruses have developed ncRNA traps to escape detection. As an example, viral-derived dsRNA induce the activation of both dsRNA dependent-PKR and TLRs, which results in type I IFN response, a well-known anti-viral defense mechanism [
26]. Different v-ncRNAs, like EBV’s Epstein-Barr virus-encoded small RNAs (EBERs), adenoviral VAI and VAII, and HIV’s trans-activation response RNA (TAR) act as a trap to inhibit PKR activation [
9]. The inhibition of IFN production is one of the most important strategies developed by viruses to block antiviral response. Indeed, it has been recently found that IFN signaling is one of the main pathway regulated by virus-derived lncRNAs. Polyadenylated nuclear RNA (PAN) from Kaposi Sarcoma Herpes virus/Human Herpesvirus-8 (KSHV/HHV8), during its lytic phase of infection, interacts with IRF4 inducing reduced expression of IFNα and IFNγ [
27]. A common feature of flavivirus is the production of viral non-coding subgenomic RNAs derived from partial degradation of the viral genome, known as subgenomic flavivirus RNAs (sfRNAs), which are involved in immune evasion. Among flaviviruses, Zika virus produces two different sfRNA, sfRNA1 and sfRNA2, which inhibit IFN production by targeting STAT2 pathway [
28,
29].
Moreover, sfRNAs produced by numerous flaviviruses, including Japanese encephalitis virus (JEV), West Nile Virus (WNV) and dengue virus antagonize the antiviral response by inhibiting the IFN signaling, the expression of IFN-β or specific IFN-stimulating genes (ISGs) [
30]. Additionally, a chimeric lncRNA HBx-LINE1, which derives from the integration of HBV into host cell genome, attenuate the IFN antiviral response inhibiting the host cell-derived miR-122 [
30]. Together with the ability to inhibit type I IFN antiviral response, v-ncRNAs are able to influence cytokines and chemokines secretion as a mechanism of immune evasion. On this purpose, ebv-miR-BART6-3p acts in two different ways: it can directly binds to RIG-1 mRNA, with consequent impaired production of the antiviral cytokine type I IFN [
31], or in association with host-derived miR-197, acts on IL-6R mRNA induced an impaired production of the pro-inflammatory cytokine IL-6 [
32]. Another EBV-derived miRNA that inhibits type I IFN signaling is ebv-miR-BART16, which has as specific target the host mRNA CREBBP [
33]. Similarly, KSHV/HHV8 produces kshv-miR-K12-11, which induces the impairment of type I IFN signaling by targeting IKKε [
34], and kshv-miR-K12-10 which reduces the production of IL-6 and IL-10 by targeting TWEAKR [
35]. Ebv-miR-BHRF-1-2-5p and ebv-miR-BART15 acts on IL-1 signaling, the former by targeting IL-1 Receptor 1 [
36] and the latter inhibiting IL-1β production [
37]. Two components of TLR/IL-1R-mediated signaling, MYD88 and IRAK1, are, respectively, the targets of the KSHV/HHV8 derived miRNAs kshv-miR-K12-5 and kshv-miR-K12-9, which affect the secretion of inflammatory cytokines [
35].
In addition, also the MHC-restricted antigen processing and presentation can be influenced by v-ncRNAs. Indeed, ebv-miR-BART2 interferes with MHC-I antigen processing because of it targets CTSB mRNA, while ebv-miR-BHRF1-3 blocks peptide transport to MHC-I, targeting TAP2 [
38], and ebv-miR-BART1-5p inhibits antigen capture and processing acting on LY75 mRNA [
39]. Both ebv-miR-BART1 and ebv-miR-BART2, linking respectively to IFI30 mRNA and LGMN mRNAs, induce the impairment of MHC-II-restricted antigen processing [
39]. Finally, an in silico analysis demonstrated that a Merkel Cell Polyomavirus (MCPyV)-derived miRNA, namely MCV-miR-M1-5p, seems to direct targeting an intrinsic antiviral protein, SP100, which leads to a reduction in the secretion of CXCL8 with a final effect of the subversion of the host-cell immune response, influencing neutrophils chemotaxis [
40].
NcRNAs can also influence host T cell behavior. In fact, two HPV-derived miRNAs, targeting different host mRNAs, such as BCL11A, CHD7, ITGAM, RAG1, and TCEA1 (miR-H1-1) or PKNOX1, SP3, XRCC4, JAK2, and FOXP1 (miR-H2), are able to inhibit T cell development and activation [
41]. Additionally, many EBV-derived miRNA target genes involved in T cell polarization and migration, namely ebv-miR-BART1, -BART2, -BART10, -BART22, and -BHRF1, prevent the polarization of CD4
+ T helper cells toward antiviral Th1 subtype acting on IL12B mRNA [
36,
38], whereas BHRF1-3 inhibits CXCL11 mRNA, blocking the chemotaxis of activated T cells [
42]. More recently, it has been suggested a role of polyomaviruses’ ncRNAs in the modified behavior of T cells during infection. Indeed, the beta-polyomaviruses John Cunningham virus (JCV), BK virus (BKV), simian virus 40 (SV40) and MCPyV encode two miRNAs (miR-S1 and miRJ1), which control the viral replication by inhibiting viral T antigen expression that lead to the suppression of antiviral T cell response [
43]. Interestingly, different HIV-1-derived ncRNAs are able to suppress Anti Sense Protein (ASP), which normally induces CD8 T cell responses during chronic infection [
44], so inhibiting CD8 T cells activation and functioning [
45]. Although the detection of v-ncRNAs in RNA viruses is controversial [
46], other than HIV retrovirus, also a negative-sense RNA virus as Ebola, is able to produce miRNA-like small RNAs [
47]. Liu et al. described as EBOV-MiR-1-5p, an analog of host miR155, inhibits the expression of importin-α5, which seems to be a potent mechanisms of immune evasion [
48].
Some v-ncRNAs are able to induce NK cell evasion of virally infected cells by inhibiting the action of NKG2D. This receptor exert its killing function by recognizing stress-induced ligands, such as MHC class I-related chain B (MICB) or UL16 binding protein 3 (ULBP3), which are upregulated on virally infected cells. In particular, it has been observed that miRNAs encoded by different herpes viruses, such as hcmv-miR-UL112-1 (CMV), ebv-miR-BART2-5p (EBV) and kshv-miR-K12-7 (KSHV), inhibit the NKG2D action by targeting MICB [
49,
50,
51]. Besides, a miRNA conserved between two different polyomaviruses (JCV and BKV), namely miR-J1-3p, targets the stress induced molecule ULBP3 [
52].
As described, different viruses, despite encoding completely different ncRNA sequences, always share common targets and mechanisms of immune evasion (
Table 1), opening an important question about the co-evolutionary development of v-ncRNAs and their matching cellular targets (
Figure 1).
4. v-ncRNA Host Mimicry
Interestingly, viruses have evolved a clever strategy to affect directly the host pathways [
91] called “mimicry”: viral miRNAs analog of host human miRNAs, with whom they share the seed sequence and potentially regulate hundreds of targets. For example, the sequence of kshv-miR-K12-11 is identical to hsa-miR-155, a host-encoded multifunctional miRNA associated with several cancers [
120]. It has been shown that kshv-miR-K12-11, like human hsa-miR-155, is capable to induce B cell expansion in mice that shows an invasive phenotype in the mice spleen [
121,
122]. Another KHSV miRNA: kshv-miR-K6-5p shares sequence similarity to the tumor-suppressive cellular hsa-miR-15/16 miRNA family [
123], resulting in an apparent nonsense mimicry, but the theory suggests the physiological role of kshv-miR-K6-5p to balance the pro-proliferative and pro-survival functions of KSHV oncogenes, negatively regulating the cell cycle [
123]. Viral miRNA expression has been also shown for two retroviruses, the simian foamy virus (SFV), and the bovine leukemia virus (BLV). In particular, sfv-miR-S4-3p mimics the sequence of cellular hsa-miR-155, while sfv-miR-S6-3p mimics miR-132. Cellular hsa-miR-155 regulate cell proliferation, on this basis one could speculate that, in this case, sfv-miR-S4-3p stimulate proliferative activity of SFV infected cells [
124]. It has been reported that another phylogenetically distant virus, the Marek′s disease virus, encodes for a miRNA, mdv1-miR-M4-5p, analog of hsa-miR-155, able to activate the oncogene c-Myc and to suppress the TGF-β signaling pathway [
125]. Also, sfv-miR-S6-3p was shown to be a functional mimic of the IFN-suppressive hsa-miR-132, thereby helping the virus escape innate immunity. Blv- miR-B4-3p mimics the sequence of hsa-miR-29, a miRNA over-expressed in lymphoproliferative disorders, suggesting that viral miRNA expression may sustain proliferation of the infected cells playing a role in BLV associated tumorigenesis [
126]. It has lately been demonstrated that a miRNA encoded by the DNA virus SV40, sv40-miR-S1-5p, has a seed sequence identical to the human hsa-miR423-5p and is able to downregulate the viral T antigen [
127]. The same study has evidenced that HIV-1-encoded hiv1-miR-N367 shared the same seed sequence to the human hsa-miR-192 targeting the same gene poly(A)-binding protein (PABP) [
127]. The examples here reported imply that this viral mechanism represents a more general phenomenon that needs to be fully unraveled.
5. Viral Circular RNA
Recently, a novel class of ncRNAs, the Circular RNAs (circRNAs), with gene regulatory functions have been discovered and investigated primarily in gammaherpesviruses [
128]. They have initially been considered as incorrect splicing products [
129] but recently a master regulatory quality has been suggested for this class of non-coding RNAs [
130]. In fact, it has been proposed they act as sponges of miRNAs [
19] and thus as regulators of gene expression at post-transcriptional level [
131]. In humans, CDR1as is the first example of regulatory circRNA that binds hsa-miR-7 preventing its binding to other molecules [
19]. Such inhibitory function is termed “sponge” function.
Several evidences indicated the existence of viral circRNAs encoded in gammaherpesvirus EBV, KSHV and murine gammaherpesvirus 68 (MHV68) [
132,
133,
134,
135]. In EBV, more than 30 circRNAs were detected, and these derived from one viral gene that encode for several circRNAs by back-splicing mechanisms [
132,
134,
135]. EBV abundantly expressed circRNAs from the BamHI A rightward transcript (BART) locus (circBARTs) were found mostly expressed during all latency programs (latency type I, type II, and type III) [
132] and across tissue and tumor types [
134]. Lymphoblastoid cell line obtained with B95-8 EBV strain, a defective EBV virus, thus lacking of miRNA and circBART expression, demonstrates they are not mandatorily required for the maintenance of the EBV genome in cell culture, but since they have been found in different tumor type and PTLD specimens, they could play important role in the viral fitness [
132]. It has been shown that one of these circRNAs, EBV cRPMS1, binds human hsa-miR-31, -miR-203, and -miR-451, leading to apoptosis as well as reduced invasiveness. Thus, cRPMS1 sponge activity plays tumorigenic functions in EBV-infected cells [
136].
Also in KSHV, at least 10 viral ORFs are reported to express circRNAs [
132,
133], some of which have oncogenic potential with promotion of cell proliferation [
133].
The circRNAs are a very interesting class of ncRNAs thanks to the fine tuning of transcription regulation that might contribute to viral oncogenesis and for this reason they will become surely subject of many studies.