Development of miRNA-Based Approaches to Explore the Interruption of Mosquito-Borne Disease Transmission

MicroRNA (miRNA or miR)-based approaches to interrupt the transmission of mosquito-borne diseases have been explored since 2005. A review of these studies and areas in which to proceed is needed. In this review, significant progress is reviewed at the level of individual miRNAs, and miRNA diversification and relevant confounders are described in detail. Current miRNA studies in mosquitoes include four steps, namely, identifying miRNAs, validating miRNA-pathogen interactions, exploring action mechanisms, and performing preapplication investigations. Notably, regarding the Plasmodium parasite, mosquito miRNAs generally bind to mosquito immunity- or development-related mRNAs, indirectly regulating Plasmodium infection; However, regarding arboviruses, mosquito miRNAs can bind to the viral genome, directly modifying viral replication. Thus, during explorations of miRNA-based approaches, researchers need select an ideal miRNA for investigation based on the mosquito species, tissue, and mosquito-borne pathogen of interest. Additionally, strategies for miRNA-based approaches differ for arboviruses and protozoan parasites.

Insecticide-based interventions [e.g., long-lasting insecticide-treated bed nets (LLINs) and indoor residual spraying (IRS)] are important components of integrated mosquito management programs designed to block the transmission of mosquito-borne diseases. The insecticides used in these interventions exert strong selection pressure on resistance and lead to the evolution and spread of mosquito resistance, representing a major concern in mosquito-borne disease control programs . The World Health Organization (WHO) claims that innovative vector control tools are urgently needed (World Health Organization, 2012). MicroRNAs (miRNAs) are single-stranded, conserved, and small endogenous noncoding RNAs that have important regulatory functions at the posttranscriptional level in diverse organisms (Bartel, 2004;Bartel, 2009). The regulation of miRNAs is indispensable for various processes, including apoptosis, development, differentiation, viral infection, and so on (Bartel, 2004;Bartel, 2009). More importantly, the functions of miRNAs can be explored and utilized. For instance, miR-15, miR-16, miR-34, and Let-7 have been patented and approved for cancer diagnosis or treatment (Mishra et al., 2016). Given their characteristics and functions, miRNAs represent one possibility for establishing a new tool, namely, miRNA-based approaches. Thus, numerous studies on mosquito miRNAs have been performed since 2005 (Wang et al., 2005) with the ultimate goal of utilizing miRNA-based approaches for disease control Feng et al., 2018a). The development of a mosquito miRNA-based approach is presumed to always follow the research roadmap of identifying mosquito miRNAs, observing miRNA-pathogen interaction, exploring action mechanisms, performing preapplication investigations and conducting clinical or field trials. Advances at each step of the roadmap need to be outlined to provide important insight into potential applications.

THE MOSQUITO MIRNA DATABASES ANALYZED
Publications focused on investigating mosquito global miRNA profiles were selected for extraction of information, including the authors, publication year, study materials, methods, miRNA names, canonical sequences, and so on. Then, this information was supplemented during the review of the other included papers. Approximately 1635 mature or predicted miRNAs were collected. Of the 1635 miRNAs, 853 (52.17%) were limited to identification and lacked any additional study information. The remaining 782 (47.83%) were further investigated; thus, they were tracked by their annotated names for study development.

STEPS OF THE RESEARCH ROADMAP
Progress in studying individual miRNAs with annotated names in the database was tracked (Tables 1-7 and S1), and an overview of study advances is provided in Figure 1. The exploration of miRNA-based approaches proceeded through the following four steps along the proposed research roadmap: identifying mosquito miRNAs (Tables 1-5); validating pathogen-miRNA interactions (Tables 6 and S1); exploring the mechanism of action, which refers mainly to target prediction and verification (Tables 7 and S1); and performing preapplication investigations (Liu P. et al., 2016). These steps involved the 20 items listed in Figure 1, for example, interactions between miRNAs and Plasmodium, dengue virus (DENV), Zika virus (ZIKA), Chikungunya virus (CHIKV), Wolbachia, West Nile virus (WNV), Palm Creek virus (PCV), Japanese Encephalitis virus (JEV), and o'nyong'nyong virus (ONNV).
However, no clinical or field trial has been reported, indicating that the miRNA-based approach may have encountered a bottleneck of application in mosquito-borne disease prevention and control, although several attempts to establish application models have been conducted (Heiss et al., 2011;Tsetsarkin et al., 2015;Tsetsarkin et al., 2016a;Tsetsarkin et al., 2016b). The details of significant progress achieved at each step are reviewed below.
More examples based on individual miRNAs are noted in Tables 2-5. Overall, the expression levels of miRNAs are regulated by complicated factors, including mosquito species, sexes, developmental stages, tissues or organs, aging, blood feeding, and so on (Tables 2-5). In addition to differences in expression levels, the preferred or functional arm also varies among these factors in terms of the change in 5p/3p ratio or even dominant arm shifts (Skalsky et al., 2010;Biryukova et al., 2014;Castellano et al., 2015). For example, the 5p/3p ratios of miR-956-3p and miR-219-5p are significantly reduced by blood feeding (Biryukova et al., 2014). Moreover, isomiR production based on acylation, uridylation, adenine and uracil extension/ addition can be induced by blood feeding and insecticide resistance (Skalsky et al., 2010;Biryukova et al., 2014).

miRNA-Pathogen Interactions in the Mosquitoes
The second step in the research roadmap always begins with an observation of statistical correlations between miRNA regulation and pathogen infection in mosquitoes. The pathogens primarily include Plasmodium, DENV, CHIKV, Wolbachia, Zika virus, WNV, JEV, PCV and ONNV ( Table 6). The miRNA abundance may vary upon pathogen infection in mosquitoes according to differences in the studied material, e.g., miR-10-5p is upregulated in CHIKV-infected Ae. aegypti (Dubey et al., 2017); conversely, it is downregulated in DENV-infected Ae. aegypti Etebari et al., 2015) (Table 6). More importantly, a few miRNAs were found to exhibit similar regulation patterns in different independent studies, and the repeatability of these results makes them more reliable (Table 6), as described in the following
As in the first study step in the roadmap, in addition to differences in expression levels, changes in 5p/3p ratio, dominant arm shifts, and isomiR production can be modified by pathogen infection (Etebari et al., 2015).
remarkably limited effects on the mosquito miRNA profile; therefore, miRNAs may not play an important role in the interaction of PCV with Ae. aegypti (Lee et al., 2017) or ONNV with Anopheles coluzzii (Carissimo et al., 2018). Thus, researchers are currently unable to select a miRNA as an ideal candidate to establish a miRNA-based approach for the control of three mosquito-borne diseases.
It is easy to find that the upregulating and downregulating miRNAs in response to pathogen infection co-exist in the mosquito ( Table 7). No miRNA has been reported to induce multiantipathogen effects on the two kinds of flaviviruses and Plasmodium protozoans. A more detailed description of the progress achieved by studies examining miRNA-pathogen interactions in mosquitoes is presented in Table S1.
In the canonical mechanism of action of miRNAs, mature miRNAs guide the RISC to the 3' untranslated regions (UTRs) of target mRNAs via complementary base pair interactions, thus regulating the expression of target genes (Hammond et al., 2000;Lee et al., 2002;Lee et al., 2003). Most miRNA-target (mRNA) interactions are consistent with the canonical action mechanism; however, exceptions have been identified for miRNA-virus interactions (e.g., DENV and CHIKV) in terms of the target type or regulatory outcome. First, during arbovirus infection, mosquito miRNAs can directly bind to the 3'-UTR of the viral genome (not necessary an mRNA), regulating virus replication Lucas et al., 2015b;Dubey et al., 2019;Yen et al., 2019), which differs from the canonical mechanism of action. However, the mechanisms underlying mosquito miRNA-Plasmodium interactions are always consistent with the canonical mechanism of action, namely, miRNAs generally bind to mosquito immunity-or development-related mRNAs, indirectly regulating pathogen infection Dennison et al., 2015;Dong et al., 2020) (Table S1). However, the exact mechanism of translational or viral repression remains unclear (Winter et al., 2007). Second, miRNAs always negatively regulate their targets by inducing mRNA cleavage (Yekta et al., 2004) or degradation (Eichhorn et al., 2014), or by repressing translation (Fabian and Sonenberg, 2012); however, positive regulation by miRNAs is repeatedly observed in mosquitoes Zhou et al., 2014;Maharaj et al., 2015;Su et al., 2019). In addition to repressing gene expression, miRNAs can also induce the expression of genes with complementary promoter sequences, switching these genes from repressed to activated Zhou et al., 2014;Maharaj et al., 2015;Su et al., 2019).
Moreover, infection with one pathogen affects coinfection with another pathogen in mosquitoes, especially for Wolbachia or engineered mosquito densoviruses (MDVs), which can modify host miRNA profiles or use a specific host miRNA to manipulate pathogen invasion in mosquitoes (Osei-Amo et al., 2012;Maharaj et al., 2015;Liu P. et al., 2016).

Preapplication Investigation
Of the 1635 putative or mature miRNAs reported in mosquitoes, only a few have advanced to the step of preapplication investigations, and the names of these miRNAs are italicized in Table S1. The first attempt to establish an application is to exploit the vector specificity and stability of MDVs, which are restricted to mosquitoes. Anti-miRNA sponges targeting endogenous let-7-5p and miR-210-3p were introduced into MDVs in Ae. aegypti (noted as AaeDV-based vectors in Figure 1), and both sponges downregulated the expression levels of these miRNAs. According to the study, this recombinant vector is useful to purposefully inhibit or promote An. funestus (Allam et al., 2016) Ae. albopictus , An. stephensi (Jain et al., 2015) -Ae. aegypti gut at 12 h PBM (Bryant et al., 2010), Ae. aegypti fat body at 6 h PBM , An. anthropophagus midguts , Ae. albopictus (Su et al., 2017) -* The "study materials" are written in two styles, namely bold and nonbold, which indicates that the miRNAs are upregulated and downregulated, respectively. "-", no evidence of upregulation or downregulation is available. & The term "specific" here indicates that miRNAs are upregulated or downregulated in one mosquito development stage when compared with the others. # The term "specific" here indicates that the miRNA is upregulated or downregulated in blood-feeding mosquitoes compared with non-blood-feeding mosquitoes, in one study , the comparisons were conducted at the time points 72 post eclosion, 6, 12, 24, 36, 48 and 72h post blood meal (PBM). $ The term "specific" here indicates that the miRNA is upregulated or downregulated in one group compared with the opposite group. DR, deltamethrin-resistant. An. gambiae (Nouzova et al., 2018) - An. gambiae (
sequences complementary to mosquito miRNAs (noted as miRNA-targeting approaches in Figure 1) into arboviruses have been established (Heiss et al., 2011;Tsetsarkin et al., 2015;Tsetsarkin et al., 2016a;Tsetsarkin et al., 2016b). The introduction of a single copy of a miRNA target sequence into the DENV genome was shown to lead to the reduction of DENV 4 replication in vivo and in vitro (Heiss et al., 2011;Tsetsarkin et al., 2015), consistent with the results of a similar study with another pathogen, JEV (Yen et al., 2013). More interestingly, multiple insertions of heterologous target sequences of different miRNAs into the virus were shown to increase virus attenuation, whereas the insertion of two or three copies of homologous sequence (the same miRNAs) into the virus did not increase virus attenuation (Tsetsarkin et al., 2016a;Tsetsarkin et al., 2016b). The two preapplication investigations indicate the possible application of miRNA-based approaches, e.g., 1) expressing a miRNA inhibitor in vector mosquitoes by establishing genetically modified mosquitoes, subsequently reducing the fitness between mosquitoes and pathogens and interrupting the transmission of mosquito-borne pathogens (Heiss et al., 2011;Tsetsarkin et al., 2015;Tsetsarkin et al., 2016a;Tsetsarkin et al., 2016b); and 2) inserting specific miRNA target sequences into the flavivirus genome, resulting in selective tissue-specific attenuation and nonhuman-range restriction of live attenuated vaccine viruses (Tsetsarkin et al., 2016a;Tsetsarkin et al., 2016b).
The study materials have ranged widely, from entire mosquitoes to the cytoplasm or nucleus of mosquito cells, from eggs to adult mosquitoes, or from sugar-fed to pathogen-infected mosquitoes (Tables 1-6).
The expression of miRNAs is regulated by complex factors, including mosquito species, sex, developmental stage, tissue or organ, age, blood feeding status, pathogen infection status and pathogen type (Tables 2-7). Thus, miRNA expression levels detected in entire mosquitoes may lead to biased results, and for one arbovirus, some miRNAs may promote infection in mosquitoes, while for another arbovirus, miRNAs may inhibit infection (Table 7). Thus, during the exploration of miRNA-based approaches for the interruption of mosquito-borne disease transmission, an irrational approach is to commonly define a miRNA as solely inhibiting or promoting pathogen infection in mosquitoes, when the actual effects of a miRNA depend on those complex factors. Most importantly, the results presented here FIGURE 2 | Responses of miRNAs to characteristics of mosquitoes or pathogens and subsequent miRNA-pathogen interactions in mosquitoes. The responses involve modifications in the expression level, isomiR production, or 5p/3p ratio or shift in dominant arm, and any of these alterations can affect the fitness between the vector and pathogen. During modification by miRNAs, the functional component is the formation of the miRNA-target RISC complex. For arboviruses, the complex can be composed of miRNAs and the virus genome (miR-RNA complex) in mosquitoes, directly regulating pathogen infection; however, for Plasmodium parasites, it is always composed of miRNAs and mosquito immunity-related mRNAs (miR-mRNA complex), indirectly regulating the infection.
indicate that the selection of a candidate miRNA according to unique conditions or objectives during miRNA-based approach development is crucial. The current statuses of individual miRNAs presented in Tables 1-7 and S1 provide guidance for selection.
As described above, the main variations in miRNAs attributed to the mosquito species or infecting pathogen include changes in the expression level, isomiR production, or 5p/3p ratio or even a shift in the dominant arm. In our opinion, these variations in miRNAs might collectively or individually affect the formulation of miRNA-target RISC complexes, and subsequently influence the fitness between the mosquito and pathogen (Winter et al., 2007). And for mosquito miRNA-arbovirus interactions, the targets of miRNAs can be RNA genomes of arboviruses, which are always mRNA obeying the canonical action mechanism. These viewpoints are presented in Figure 2.
Moreover, although the canonical action mechanism of miRNAs always results in repression, the mosquito miRNAtarget interaction can lead to two possible forms of regulation, namely, repression or enhancement of pathogen infection in mosquitoes. Both upregulation and downregulation of miRNAs in response to pathogen infection widely coexist in the mosquito, subsequently promoting and inhibiting pathogen infection, respectively. In our opinion, these findings suggest that inhibitory and inducing miRNA expressions are essential to balance the miRNA-pathogen interaction, maintaining persistent infection and preventing considerable harm to the mosquito (Figure 3).
Currently, the antiviral effects of mosquito miRNAs on pathogens in combination with genetic engineering and molecular biology techniques may allow the use of these miRNAbased approaches as new tools to interrupt the transmission of mosquito-borne diseases. In this review, the significant progress achieved at the level of individual miRNAs facilitates the selection of an abundant, specific and effective mosquito miRNA (see Tables 1-7 and S1) that can be referenced for further research with different and specific objectives to increase the pace of development of applications and overcome the bottleneck (Tsetsarkin et al., 2015). More importantly, mosquito miRNAs can directly bind to the arbovirus genome, modifying viral replication. However, regarding the Plasmodium parasite, mosquito miRNAs generally bind to mosquito immunity or development-related mRNAs, indirectly regulating Plasmodium infection. Hence, the strategies for miRNA-based approaches differ for arboviruses and protozoan parasites.

AUTHOR CONTRIBUTIONS
Conceptualization and formal analysis: T-LX. Data curation: T-LX, Y-WS, and X-YF.
Supervision: BZ and X-NZ. Writing-original draft: T-LX. Writing-review & editing: BZ and X-NZ. All authors contributed to the article and approved the submitted version.

FUNDING
BZ received a grant from the National Science and Technology Major Program of China (No. 2018ZX10734-404). This project was financially supported by Ministry of Science and Technology of the People's Republic of China. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.