Harnessing endogenous miRNA targeting ZIKV: A cutting-edge strategy to inhibit virus infection

Abstract

Emerging RNA virus outbreaks, including Zika virus, highlight the urgent need for novel antiviral strategies. Zika virus, a positive-strand RNA virus, causes congenital Zika syndrome, and to date, there are no approved vaccines or antiviral treatments. In this context, microRNAs are small non-coding RNAs that regulate gene expression and show potential as antiviral agents due to their ability to target viral RNA, making them a promising therapeutic approach against Zika syndrome. In this study, we identified endogenous microRNAs that interact with the virus genome using computational algorithms and overexpressed them in VERO cells. Twelve microRNAs reduced viral cytopathic effects by more than 50% in cells infected with a Brazilian Zika virus strain. Additionally, we used a computational platform to select pharmacological compounds capable of modulating endogenous microRNAs in human cells, achieving over 90% inhibition of Zika virus activity. These findings offer a promising path through drug repurposing for antiviral therapy by modulating endogenous microRNAs, with potential applications for other positive-strand RNA viruses.

Graphical abstract

This study explores endogenous microRNA modulation as an antiviral strategy against Zika virus. By identifying key miRNAs and leveraging drug repurposing, Silveira Martins and colleagues achieved significant viral inhibition. These findings open new avenues for virus therapeutics, demonstrating the potential of modulating host miRNA regulatory pathways to inhibit viral infections.

Introduction

The Zika virus (ZIKV) is an arbovirus belonging to the genus Flavivirus and family Flaviviridae. It is characterized as an enveloped, non-segmented, positive single-stranded RNA virus that encodes a polyprotein comprising three structural proteins (C, prM, and E) and seven non-structural proteins (NS1, NS2a, NS2b, NS3, NS4a, NS4b, and NS5).^1,2,3,4 ZIKV is primarily transmitted by infected Aedes mosquitoes and can also be transmitted through sexual contact and during the perinatal period.^5,6,7 Most infections are asymptomatic; however, in pregnant women, it may lead to severe malformations described as congenital Zika syndrome (CZS).^8,9,10,11 The virus remains stable in human fluids such as saliva, urine, semen, and breast milk for weeks to months, and it has been reported that in infected men, the virus can reside in the testes for long periods, potentially leading to infertility.^12,13,14 Despite the severe problems associated with the Zika virus, there are still no effective vaccines or antiviral treatments available.¹⁵ The increasing emergence and re-emergence of infectious outbreaks like ZIKV and other RNA viruses highlight the need for the development of innovative antiviral approaches.

In this way, microRNAs (miRNAs) can emerge like good antiviral candidates.^16,17,18,19 miRNAs are small non-coding RNA molecules approximately 22 nucleotides in length that are synthesized by cells and play an important role modulating gene expression.²⁰ Their synthesis follows the canonical model, in which non-coding regions are transcribed by RNA polymerase II, generating pri-miRNAs. These are processed by the Drosha complex in the nucleus and by the Dicer enzyme in the cytoplasm, resulting in mature miRNA molecules that are incorporated into the RISC complex to drive the degradation or repression of target messenger RNAs (mRNAs).^20,21 The currently known database lists over 2,000 human miRNA precursors with the potential to regulate thousands of targets, including viral transcripts.²²

Due to structural and processing similarities between viral RNA (vRNA) and messenger RNA (mRNA) molecules, miRNAs have the potential to target vRNA regions and antagonize viral progression by inactivating their replication and/or translation of viral proteins.²³ Studies have shown that exogenous miRNA molecules can enter cells and exert antiviral effects, categorizing them as a potential new class of antivirals.^24,25,26,27 Because of their action through a small portion of base complementarity and the numerous in silico tools for predicting candidates, this class of miRNA-based antivirals is considered flexible and offers the advantage of providing a rapid response during possible new infectious outbreaks.^22,28,29 However, despite the success of Miravirsen (a synthetic antisense oligonucleotide used to treat hepatitis C) at clinical trial, studies point out that the use of synthetic miRNAs or anti-miRNAs may have implications, such as toxicity and off-target effects.^30,31

In the context of ZIKV, many studies have been conducted to evaluate the expression of miRNAs associated with viral infection.^32,33,34,35 However, few studies have explored the use of these miRNA molecules as a therapeutic strategy. Furthermore, while many studies use synthetic miRNAs,^36,37,38,39 here we explore the modulation of their endogenous expression. In this regard, our work explored a combination of in silico screening of virus-targeting miRNAs and a platform to find candidate drugs to enhance their endogenous expression. By this mean, enhancing endogenous defenses by miRNA modulation may lead to the development of protective approaches and treatments to ZIKV infection.

Results

Prediction of miRNAs with potential interaction on ZIKV genome

Since our hypothesis was that ZIKV could be inhibited by miRNA, the first step was to discover potential human endogenous miRNA candidates that could inactivate viral RNA. For this, we used the PITA target prediction algorithm,⁴⁰ to search for miRNA target sites within the ZIKV genome, identifying 2,589 mature human miRNA molecules with potential binding sites in the ZIKV vRNA genome. The miRNA candidates were ranked based on a score that considered target site accessibility, free energy of the interaction, and the thermal stability of the interaction. This scoring system prioritized sequence complementarity between miRNA candidates and the viral genome, assigning more negative scores for best candidates (Table S1). Besides miRNA score, another criterion that was observed to investigate antiviral effect was to choose candidates that specifically interacted with different portions of the viral genome, as structural and non-structural regions. In this way, once we have the PITA’s prediction for the best 100 miRNA target sequences, we manually choose candidates that exhibited interaction with the mature miRNA seed region (7-mer) in the final fifth of the coding sequence of each viral protein. Two to three candidates were selected for each coding region distributed across the viral genome, and selected miRNAs are listed in Table 1.

Table 1.

miRNA candidates selected from the region of interaction with coding portions of viral proteins

ZIKV coding regions	miRNA candidates
Capsid (C)	hsa-miR-4722-5p; hsa-miR-6839-3p
prM	hsa-miR-7703; hsa-miR-1301-3p; hsa-miR-939-3p
Envelop (E)	hsa-miR-4786-3p; hsa-miR-3147; hsa-miR-7974; hsa-miR-6769b-5p; hsa-miR-320a-3p
NS1	hsa-miR-6855-5p; hsa-miR-6764-5p; hsa-miR-1238-5p
NS2A	hsa-miR-6783-3p; hsa-miR-4659b-3p; hsa-miR-4640-5p
NS2B	hsa-miR-7110-3p; hsa-miR-4691-5p; hsa-miR-3138
NS3	hsa-miR-3180-5p; hsa-miR-6830-3p; hsa-miR-1227-3p
NS4A	hsa-miR-4776-3p; hsa-miR-8089; hsa-miR-7158-5p
NS4B	hsa-miR-6749-5p; hsa-miR-1244
NS5	hsa-miR-4763-3p; hsa-miR-1909-5p; hsa-miR-4726-3p

ZIKV infection is inhibited by miRNA expression targeting viral genome

To investigate whether the miRNA candidates had the potential to inhibit ZIKV infection, we performed a cell culture assay to evaluate ZIKV inhibition by miRNA. In this assay, VERO cells were genetically modified with lentiviral vectors to overexpress the previously chosen miRNA candidates that could target the viral genome. The VERO-miRNA cell lines were then challenged with different concentrations of ZIKV, and infection was assessed using a TCID₅₀ assay. We observed that some of the miRNA candidates were able to significantly reduce the cytopathic effect by more than 50% compared to the control parental VERO cell line that has no human miRNA (Figure 1).

Figure 1.

Expression of miRNA candidates that show in silico interaction with the ZIKV genome reduces tissue cytotoxicity in infected cells

(A) Representative diagram of the selected miRNA candidates in relation to the complementary coding regions E, prM, C, NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 of the ZIKV genome. (B) Cell lines overexpressing miRNA candidates were infected with ZIKV using serial dilutions from 10⁻¹ to 10⁻¹⁰ to assess the cytoprotective potential of the miRNA candidates of interest. The cells were cultured for 6 days post-infection, followed by blind observation and counting of the wells displaying cytopathic effects under an optical microscope. Tissue damage values per mL for each analyzed cell line were normalized to the values observed in parental cells (black column). Results are expressed as arithmetic mean +/− standard error of the mean. ∗∗∗p < 0.0001 and ∗∗p < 0.001 relative to the control group of parental cells (black column).

In silico analysis to find miRNA modulatory drug candidates

Once the cell culture assay using VERO cells overexpressing miRNA explored the potential to inhibit virus infection, our next challenge was to discover how to enhance endogenous miRNA expression, to inhibit virus infection. In this way, we used an in silico platform called METACORE (Clarivate Analytics), which allowed us to understand how these endogenous miRNAs could be modulated by drugs to increase their endogenous expression in the cells of interest, with the goal of preventing virus infection. Given the limited studies correlating possible pharmacological compounds with the modulation of the chosen miRNA candidates’ expression, we considered that the sequences of these miRNAs are present in non-coding regions, particularly introns.^{41,42,43,44,45} Therefore, by promoting the increased synthesis of the proteins encoded by genes harboring these introns, we could, as consequence, boost the expression of desired miRNAs. By generating networks indicating which molecular pathways could modulate the expression level of target proteins, we identified three pharmacological compounds, Berberine, Prednisolone, and Tetracycline. These molecules modulate 11 ZIKV-targeting miRNAs and consequently could inhibit virus infection (Table 2).

Table 2.

Dataset of information related to the selected miRNA candidates using METACORE bioinformatics platform

miRNA	Protein	Activation pathways	Drug compound
hsa-miR-4722-5p	PIEZO/FAM38A	AHR→KLF17→PIEZO/FAM38A46,47,48,49,50	Berberine
hsa-miR-939-3p	CPSF1	AHR→ZGLP1→CPSFL150	Berberine
hsa-miR-1301-3p	DNMT3A	AHR→SMAD146,47,49,51	Berberine
hsa-miR-4786-3p	NDUFA10	AHR→FOXP3→NDUFA1052	Berberine
hsa-miR-4659b-3p	AGPAT5/PCLE	AHR→FOXP3→PLCE52	Berberine
		AHR→KLF17→PLCE53,54
hsa-miR-6783-3p	KIF18B	AHR→SMAD1→MCAK→KIF18B55,56	Berberine
hsa-miR-4722-5p	PIEZO/FAM38A	GCR→KLF17→FAM38A53,54	Prednisolone
hsa-miR-1303-3p	DNMT3A	GCR→GLIS3→DNMT3A57	Prednisolone
hsa-miR-4659b-3p	AGPAT5	GCR→KLF17→PCLE53,54	Prednisolone
hsa-miR-1909-5p	REXO1	GCR→GLIS3→REXO157	Prednisolone
hsa-miR-4786-3p	NDUFA10	PXR/RXR-alpha→coREST complex/LSD1→NDUFA1058,59	Tetracycline

The candidate compounds were chosen considering those capable of stimulating more than one miRNA candidate, leading to a better overall antiviral effect. Additionally, we looked for compounds that exhibited an activation profile with fewer interactions with other molecules, to minimize adverse effects that could disrupt cellular homeostasis during treatment and so reduce toxicity. Last, we prioritized testing compounds that are available to the population and are already approved by regulatory agencies, then repurposing for a new therapeutic application.

Evaluation of antiviral effects of the drug candidates

To evaluate the antiviral activity of the identified pharmacological compounds by in silico analysis, we performed an in vitro assay treating cells that were challenged with ZIKV. Four different cell lines, which were susceptible to ZIKV infection, were assessed for viral titer as well as drug toxicity. Treatments with Berberine, Prednisolone, and Tetracycline resulted in significant reductions in viral titers. The antiviral effect of the drugs was correlated with the modulation of miRNAs targeting the inhibition of the ZIKV genome.

To confirm whether viral inhibition was associated with the modulation of endogenous miRNA, we selected specific treatment conditions to analyze miRNAs expression using qPCR. Treatment with Berberine reduced viral titer up to 100% in the MCF7 cell line at concentrations of 10, 20, and 40 μM, indicating a strong antiviral potential (Figure 2). Based on the data from our in silico study, cell culture assays were performed, and it could be observed that the drug significantly modulated the expression of hsa-miR-939-3p, hsa-miR-4659b-3p, hsa-miR-4722-5p, and hsa-miR-6783-3p, with an induction of approximately 18- to 60-fold compared to the untreated controls (Figure 3).

Figure 2.

In vitro assay to verify inhibition of viral infection in cell cultures treated with Berberine, Prednisolone, and Tetracycline

For viral titration human cell lines (A549, HT1080, HUH7, and MCF7) were treated with Berberine at concentrations of 5, 10, 20, and 40 μM; Prednisolone at concentrations of 0.5, 1,1.5, 2, 2.5 mg/mL; and Tetracycline at concentrations of 0.01, 0.05, 0.1, 0.15, 0.2 mg/mL for 16 h, infected with ZIKV MOI 5, and maintained in culture and treatment for 48 hpi. After 48 hpi, the supernatant was collected and sent for titration experiments in VERO cells. NT represents the group of supernatants collected from infected but non-treated cells. Results are expressed as the arithmetic mean +/− standard error of the mean. ∗∗∗∗p < 0.0001 and ∗∗∗p < 0.001 relative to the non-treated control group (first column). For cell viability, the same human cell lines were treated with Berberine, Prednisolone, and Tetracycline using the same concentrations for 64 h, treated with MTT for 1 h, and analyzed for cell viability using a spectrometer. NT represents the non-treated cell group and PC the positive control. Results are expressed as the arithmetic mean +/− standard error of the mean. ∗∗∗∗p < 0.0001 and ∗∗∗p < 0.001 relative to the non-treated control group (first column).

Figure 3.

Positive modulation of miRNA candidates after treatment with Berberine, Tetracycline, and Prednisolone

The MCF7 cell line was treated with Berberine at a concentration of 20 μM for 64 h. HT1080 cell line was treated with Prednisolone at concentrations of 0.5, 1, and 1.5 mg/mL for 64 h. A549 cell line was treated with Tetracycline at concentrations of 0.1, 0.2, and 0.4 mg/mL for 64 h. At the end of this period, the cell pellets were collected and total miRNAs and total RNA were extracted, followed by cDNA synthesis and qualitative quantification by q-PCR of the expression of the candidates of interest compared to the control group. NT represents the untreated cell group. Results are expressed as arithmetic mean +/− standard error of the mean. ∗p < 0.005 relative to the untreated control group.

Prednisolone treatment reduced the viral titer by more than 90% in HT1080 cells (Figure 2). The in silico study suggested that Prednisolone could modulate hsa-miR-4659b-3p and hsa-miR-4722-5p, showing approximately 100-fold induction in the qPCR assay (Figure 3).

When investigating the treatment with Tetracycline, which reduced viral titers by up to 100% in A549 cells (Figure 2), we observed that the expression of hsa-miR-4786-3p was modulated with a 10-fold increase compared to the control (Figure 3).

The inhibition of miRNA molecules targeting the viral genome rescues ZIKV infection

To confirm that virus inhibition was associated with the endogenous modulation of miRNAs targeting the ZIKV genome, we established cell lines using viral vectors encoding antimiR cassettes to silence the ZIKV-inhibitory miRNAs. This approach, known as miRNA sponge,⁶⁰ allowed us to investigate the pharmacological effect by comparing the antimiR cell line with the parental cell line under drug treatment stimulation.

Since our in silico study suggested that hsa-miR-4786-3p was the only miRNA modulated by Tetracycline, we established an anti-miR-4786-3p A549 cell line that continuously expresses antimiR molecules targeting hsa-miR-4786-3p. When challenged with ZIKV, this antimiR cell line exhibited a rescue of viral infection by approximately 40% (Figure 4). The in silico data also suggested that Prednisolone could modulate the expression of four miRNAs targeting the ZIKV genome. We selected hsa-miR-4722-5p and hsa-miR-4659b-3p for comparative assays, which were performed in HT1080 cell lines transduced with lentivectors harboring antimiR against these targets. We found that miRNA inhibition in Prednisolone-treated cells led to a 60% rescue of virus titer when hsa-miR-4722-5p was inhibited and approximately 20% when hsa-miR-4659b-3p was inhibited through their antimiR molecules (Figure 4).

Figure 4.

AntimiR rescues ZIKV infectivity by blocking miRNA molecules

MCF7 and MCF7 antimiR cell lines for hsa-miR-4722-5p and hsa-miR-4659b-3p were treated with Berberine at concentrations of 5, 10, 20, and 40 μM for 16 h, infected with ZIKV MOI 5, and maintained in culture and treatment for 48 hpi. After 48 hpi, the supernatant was collected and sent for titration experiments in VERO cells. HT1080 and HT1080 antimiR cell lines for hsa-miR-4722-5p and hsa-miR-4659b-3p were treated with Prednisolone at concentrations of 0.5 and 1 mg/mL for 16 h, infected with ZIKV MOI 5, and maintained in culture and treatment for 48 hpi. After 48 hpi, the supernatant was collected and sent for titration experiments in VERO cells. A549 and A549 antimiR cell lines for hsa-miR-4786-3p were treated with Tetracycline at concentrations of 0.05, 0.1, and 0.15 mg/mL for 16 h, infected with ZIKV MOI 5, and maintained in culture and treatment for 48 hpi. After 48 hpi, the supernatant was collected and sent for titration experiments in VERO cells. NT represents the group of supernatants collected from infected but non-treated cells. Results are expressed as the normalized arithmetic mean +/− standard error of the mean. ∗∗∗∗p < 0.0001 and ∗∗∗p < 0.001 relative to the non-treated control group.

For Berberine treatment, which also modulates miRNAs hsa-miR-4722-5p and hsa-miR-4659b-3p, like Prednisolone, we performed assays with the MCF7 cell line. However, we observed only a rescue of 5% on virus expression when antimiR molecules for these candidates were expressed. This lower level of rescue could be explained by the presence of other miRNAs that are also modulated by Berberine, which were not targeted by the antimiRs (Figure 4).

Discussion

Our study explores a pioneering strategy to counteract ZIKV progression by endogenously stimulating specific miRNA candidates. We developed an approach that combines in silico and in vitro platforms to identify miRNA candidates targeting the Zika virus genome, which can be endogenously upregulated in target cells by treating with pharmacological compounds. Using this innovative approach, we confirmed that Berberine, Prednisolone, and Tetracycline modulate the miRNAs hsa-miR-4722-5p, hsa-miR-4659b-3p, and hsa-miR-4786-3p, effectively blocking ZIKV replication in different cell lines.

Several studies have investigated the correlation between miRNA expression changes during ZIKV infection^32,33,34,35 but few, if any, have explored these miRNA expression pathways as an alternative to fight the infection mechanism. In this work, we employed bioinformatics tools to identify miRNAs that interact with specific regions of the ZIKV genome, revealing that more than 2,000 mature miRNA molecules target the viral genome (Table S1). These findings suggest a vast collection of molecules that could be harnessed to reduce ZIKV or even other positive-sense RNA virus replication. We refined these results by focusing on miRNAs complementary to the coding sequences at the terminal regions of viral proteins, as the literature suggests that interactions with the 3′ UTR are more effective in inhibiting mRNA translation.⁶¹

By exploring a different approach from the commonly used synthetic miRNA treatments,^36,37,38,39 we worked with genetically modified cells to express the miRNA candidates, employing a model that more closely resembled physiological modulation. This approach mitigated common issues associated with synthetic miRNAs, such as instability, toxicity, off-target effects, and concentration doses, that exceed ideal physiological conditions.^31,62,63,64 In our study, we observed that 12 miRNAs—hsa-miR-4722-5p, hsa-miR-939-3p, hsa-miR-1301-3p, hsa-miR-320a-3p, hsa-miR-4786-3p, hsa-miR-3147, hsa-miR-1238-5p, hsa-miR-6855-5p, hsa-miR-6783-5p, hsa-miR-4659b-3p, hsa-miR-7178, and hsa-miR-1909-5p—reduced the cytopathic effects of ZIKV infection, indicating their antiviral potential (Figure 1).

When exploring endogenous miRNA modulation as a therapeutic strategy, we observed that half of the miRNAs are processed from intronic regions.^{41,42,43,44,45,65} Therefore, increasing the endogenous expression of these molecules would require modulating the proteins within which their genomic sequences are embedded. The study of protein modulation has been widely explored and can be easily performed using the in silico METACORE software. Metacore data are derived from previously conducted experiments and analyzed literature, enhancing their translational predictive value. We employed this platform to identify pharmacological compounds that modulate these pathways through neural networks of protein activation, selecting drug candidates that interact directly or within two nodes of the relevant pathways and avoiding pathways with significant secondary activities to mitigate cytotoxic effects (Table 2). Using this approach, we identified three pharmacological candidates capable of modulating the expression of the miRNAs of interest.

Berberine is a natural alkaloid extracted from several species of medicinal plants and has demonstrated broad therapeutic potential.⁶⁶ Studies have shown its antiviral effects in treating infections such as Chikungunya,⁶⁷ Herpes Simplex,⁶⁸ Papilloma virus,⁶⁹ Zika virus,⁷⁰ and Dengue virus.⁷¹ In our study, Berberine significantly decreased viral titers in different cell lines (Figure 2), correlating with increased expression of hsa-miR-939-3p, hsa-miR-4659b-3p, hsa-miR-4722-5p, and hsa-miR-6783-3p (Figure 3). Additionally, we observed a rescue of viral titers when cells expressing an antimiR system reduced the availability of hsa-miR-4659b-3p and hsa-miR-4722-5p (Figure 4), suggesting their involvement in the antiviral effects of Berberine. Thus, our findings demonstrate that one potential mechanism by which Berberine exerts its antiviral activity against ZIKV is through the modulation of miRNA molecules, a mechanism that has been largely unexplored in literature.

Prednisolone belongs to the class of glucocorticoids, commonly used to treat inflammatory conditions and autoimmune diseases. Recent studies have demonstrated its benefits in the treatment of severe viral infections.^72,73 In our study, Prednisolone was predicted to interact and modulate the nuclear receptor GCR (Table 2; Figure S2), which, in turn, could influence other pathways (KLF17 and GLIS3) and increase the expression of four miRNA candidates. Our results showed that Prednisolone treatment significantly reduced viral titers in HT1080 cells (Figure 2) and potentially correlated with the increased expression of hsa-miR-4722-5p and hsa-miR-4659b-3p after treatment (Figure 3). Furthermore, blocking hsa-miR-4722-5p and hsa-miR-4659b-3p in Prednisolone-treated HT1080 antimiR cells resulted in a significant rescue of viral titers, supporting the antiviral modulatory potential of these miRNAs in response to ZIKV infection using the compound.

Tetracycline is a broad-spectrum antibiotic with therapeutic applications described for over 50 years, demonstrating efficacy in both bacterial and viral infections. Its use has shown inhibitory effects on Dengue virus^74,75 and Zika virus.⁷⁶ Similarly, our study (Figure 2) demonstrated that treatment with Tetracycline produced satisfactory inhibitory effects in three out of four cell lines analyzed. In particular, the A549 cell line showed approximately 90% reduction in viral titer. The antiviral effect of Tetracycline appears to be related to the modulation of hsa-miR-4786-3p, which showed a 10-fold increase in expression in treated cells (Figure 3). Moreover, blocking hsa-miR-4786-3p with an antimiR system led to a 40% rescue of viral titers (Figure 4), indicating that this miRNA may play a role in Tetracycline’s mechanism of action in ZIKV-infected cells. Although the inhibitory effect of Tetracycline derivatives on ZIKV has been previously documented,⁷⁶ the precise mechanisms remain unclear. Based on our findings, we propose that viral inhibition may occur by the upregulation of miRNA candidates that inactivate the viral genome, offering a potential mechanism for Tetracycline’s antiviral activity.

Comparing the results of Berberine, Prednisolone, and Tetracycline, we observed that the more pathways and molecules need to be modulated, the more complex the biodistribution network becomes, potentially reducing the overall effect. Berberine, which modulates the AHR pathway, and Tetracycline, which modulates the PXR/RXR-alpha-coREST complex/LSD1 pathway, both demonstrated better inhibitory effects across the cell lines studied, likely due to their broader cellular biodistribution profiles. In contrast, Prednisolone, which modulates GCR with a more specific distribution (KLF17 or GLIS3), exhibited greater variability in inhibitory efficiency among the analyzed cell lines. Although off-target effects are a common concern in therapies—arising from both the bioavailability of compounds and the miRNA molecules—variations in the transcriptomes of target cells present opportunities for refinement. Targeting miRNA molecules expressed in specific microenvironments where the virus acts may help reduce this bias. Platforms such as TissueAtlas, GTEx Portal, and UCSC Genome Browser can be used to explore these variations in expression. As demonstrated in our studies and supported by literature, miRNA expression is tissue-specific.^77,78 Thus, the modulation of miRNA candidates showed different behaviors across the cell lines tested, impacting on the final antiviral effects, which could be leveraged to enhance the efficiency of therapeutic treatments.

The three compounds analyzed in our study have already been approved for use by regulatory agencies and have established safety profiles, making them promising candidates for pharmacological repurposing. Although these compounds have usage restrictions for pregnant women,^79,80,81,82 they could be considered for treating other severe conditions associated with ZIKV infection, such as male infertility and Guillain-Barré syndrome. Additionally, further studies could assess the safety profiles of these compounds for use in pregnant women, potentially through chemical reformulation.^83,84 Notably, Berberine has been investigated for its potential to reverse fertility issues, suggesting that its application during the first trimester of pregnancy warrants further investigation and reevaluation.⁸⁵

Despite the promising results achieved with our methodologies, we acknowledge the need for future RNA sequencing and in vivo experiments to assess the safety concerns associated with endogenous miRNA modulation through compound treatment. These studies would provide further validation of the pharmacological candidates identified in our analysis, ensuring their efficacy and safety for therapeutic applications, including their potential repurposing for ZIKV-associated conditions.

Collectively, our findings highlight the potential for efficiently identifying miRNA candidates that interact with the viral genome through the application of bioinformatics tools. By adopting a rational approach that combines algorithmic predictions with the distribution of these interactions across coding regions, we expanded the scope of positive results. Additionally, the analysis of pathways and identification of potential endogenous modulators via treatment with pharmacological compounds—predicted by the METACORE platform—accelerated and reinforced the outcomes of our experiments. We introduced a novel strategy for controlling viral replication by modulating endogenous miRNAs to enhance the body’s natural defenses against infections. Our findings represent a promising and innovative approach for treating ZIKV infections, offering a viable alternative to conventional antiviral therapies. Moreover, these results may pave the way for future research and the development of miRNA-based therapies for other significant infections, while also advancing our understanding of miRNA-virus interactions, particularly in the context of Zika virus infections.

Materials and methods

Virus

The Brazilian ZIKV strain BeH823339 used in the experiments was kindly provided by Dra. Clarice Weis Arns (University of Campinas). To expand virus stock, VERO cells were infected with the virus sample diluted in 2 mL of DMEM medium (Gibco), supplemented with 2% fetal bovine serum (FBS) (Vitrocell) and incubated for 2 h at 37°C and 5% CO₂. The supernatant was discarded and added 7 mL of DMEM 2% FBS culture medium, following 72 h of incubation at 37°C, 5% CO₂. Next, cells were lysed performing three cycles of freezing and thawing to release virus particles. The suspension was then centrifuged at 1,000 g for 5 min at 4°C, transferred to a cryotube, and stored at −80°C. ZIKV was titrated by plaque assay protocol.

Cell lines

VERO cell lines (ATCC CCL-81), A549 (ATCC CCL-185), HT1080 (ATCC CCL-121), HUH7 (ECACC 01042712), and MCF7 (ATCC HTB-22) were cultivated in DMEM medium (Gibco) supplemented with glutamine and 10% FBS (Vitrocell), at 37°C and 5% CO₂.

Reagents

Both compounds Tetracycline (C₂₂H₂₄N₂O₈ · HCl) and Berberine (C20H18ClNO4 · xH2O) (Sigma-Aldrich) were diluted in sterile Milli-Q water, Tetracycline was kept at 100 mg/mL, and Berberine was kept in a stock concentration of 2 mM. The Prednisolone sodium phosphate stock (PredSim—Mantecorp Farmasa) was provided at 11 mg/mL.

In silico prediction of miRNA candidates using PITA algorithm

The miRNA candidates were identified by PITA algorithm (Weizmann Institute, Israel),⁴⁰ which was able to search human target miRNA sites within the ZIKV genome, from Brazilian variant BeH823339, deposited in GenBank. The software parameters were set to find candidates with a high complementarity to the sequence of virus genome, harboring at least a 7 bases seed. These candidates were ranked based on the number of interactions with the viral genome and the thermostability of these interactions. The list of candidates was manually reviewed to select those complementary to the ZIKV genome distributed in different regions of structural or non-structural genes as E, prM, C, NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5 sequences.

Establishment of VERO-miR lineages

To establish cells that could overexpress the miRNA candidates, we transduced VERO cells with lentivectors harboring miRNA candidates. Lentiviral vector plasmids were generated by cloning a PCR fragment amplifyed from gDNA of human 293T cells, using Promega kit (M3001) following manufacturer protocol. We designed oligonucleotides to amplify the fragment harboring the pri-miRNA (Table S2). The PCR product was purified and digested with EcoRI enzyme (NEB #R3101) and cloned into the FUGW lentiviral vector (Addgene#14883) downstream of the GFP reporter gene. The lentiviral vectors were generated by the Viral Vector Laboratory LVV/LNBio as previously described.^86,87,88 VERO cells were transduced with lentivirus using a multiplicity of infection (MOI) of 10, supplemented with 8 μg/mL of polybrene, and expanded.

Evaluation of antiviral activity of miRNA candidates using VERO-miR cells

The antiviral activity of miRNA candidates was evaluated by cell culture assays comparing the results of a ZIKV infection challenge of VERO-miR cells (that overexpress human miRNA candidates) to parental VERO cells (negative control that have no human miRNA). The TCID₅₀ method was used to evaluate infection, following Spearman–Kärber method.⁸⁹

In silico assay prediction of drug compounds that modulates miRNA candidates

The prediction of pharmacological compounds that could modulate miRNA expression in human cells was given by the METACORE platform (Clarivate Analytics). The software generates a wide network range, suggesting interactions among molecular targets. The network of proteins encoding each miRNA candidate gene in their intronic sequence portion were generated for analysis. These networks were examined by investigating pathways that would interact within two nodes of distance. Once the interacting pathways were identified, it was searched for possible compounds that could modulate these pathways.

Evaluating antiviral activity of drug-induced miRNA candidates

We performed in vitro assays using A549, HT1080, HUH7, and MCF7 cells that were infected with ZIKV. The procedure consisted of two steps. First, human cells were infected and treated with the selected compounds. Second, the supernatants were harvested at the end of the incubation/treatment period and titrated by plaque assay in VERO cells. The cells were seeded in 24-well plates the day before infection, and after cells adhered to the plate, culture medium was replaced, and the compounds were added to initiate pre-treatment with the indicated drug concentration, as 5, 10, 20, and 40 μM for Berberine; 0.05, 0.1, 0.15, and 0.2 mg/mL for Tetracycline; and 0.5, 1, 1.5, 2, and 2.5 mg/mL for Prednisolone. The next day, cells were infected with ZIKV using an MOI of 5 and incubated 1 h at 37°C, 5% CO₂. Thereafter the culture medium was replaced, and cells were incubated for an additional 48 h at 37°C, 5% CO₂. The supernatant from each experimental condition was harvested 48 h later and titrated by plaque assay. Concentration of Tetracycline and Prednisolone was expressed in mg/mL, as described in in vitro studies in the literature.^90,91 On the other hand, the concentration of Berberine was expressed in micromolar, as the IC50 values obtained for this substance fall within this order of magnitude.

Plaque assay

VERO cells were seeded in a 24-well plate at 2 × 10⁵ cells/well. The ZIKV infection was done 24 h later, using 250 μL of serial dilutions of virus supernatants, incubating 2 h at 37°C, 5% CO₂. After infection, the supernatant was replaced by DMEM medium supplemented with 2% CMC (carboxymethylcellulose), 2.5% FBS. These cells were fixed 7 days later, in 10% paraformaldehyde, and stained with 1% violet crystal (#V5265 Sigma). The plaques were counted to estimate virus titers.

Viability assay

Human cell lines (A549, HT1080, HUH7, and MCF7) were seeded in 96-well plates. The cells were treated with Berberine (5, 10, 20, and 40 μM); Tetracycline (0.05, 0.1, 0.15, and 0.2 mg/mL) or Prednisolone (0.5, 1, 1.5, 2, and 2.5 mg/mL) for 64 h. After this incubation period, the culture medium was replaced by 100 μL of fresh DMEM medium supplemented with 10 μL of MTT per well. The cells were incubated for an additional hour at 37°C, 5% CO₂. The supernatant was removed, adding 75 μL of DMSO, followed by additional 10-min incubation. The cells were evaluated for viability using the CLARIOstar equipment at 540 nm.

miRNA isolation and qPCR

Drug-induced miRNA expression was evaluated by qPCR. Oligonucleotide primers were designed based on the stem-loop RT-PCR protocol, previously described by Varkonyi-Gasic et al.,⁸³ available in supplementary materials. The miRNA fragments were isolated from the cell pellets by the RNeasy MinElute Cleanup kit (Cat. No./ID: 74204) from Qiagen. After extraction, the miRNAs were quantified, using 2 μg of sample to the hybridization step. Samples were incubated in a heat block at 65°C for 10 min and 1 min on ice. After hybridization, cDNA synthesis was performed using the High-Capacity cDNA Reverse Transcription Kit (43-688-14) from Applied Biosystems, and the products from each condition were amplified using SYBR Green Master Mix (1725121) from BioRad following manufacturer’s protocol.

Establishment of antimiR-harboring cell lineages

The cell lines were generated by transduction with lentiviral vectors harboring antimiR cassettes, encoding hsa-miR-4722-5p, hsa-miR-4659b-3p, and hsa-miR-4786-3p. The lentivirus preparations were generated by the Viral Vector Laboratory at CNPEM as previously described.^86,87,88 Plasmid transfer vectors PLKO.1-TRC were generated by cloning DNA oligonucleotides encoding antimiR sequences (Table S4). These oligos were annealed, then the duplex was digested with the AgeI and EcoRI enzyme (NEB), and cloned into the pLKO.1-TRC vector (Addgene#10878). To establish permissive cells harboring the antimiR of interest, we transduced A549 (ATCC CCL-185), HT1080 (ATCC CCL-121), and MCF7 (ATCC HTB-22) with the pLKO.1-antimiR lentiviral vectors, using an MOI of 10, and expansion was performed as the cells reached confluence after treatment with puromycin.

Flow cytometry

Flow cytometry assays were performed using a NovoCyte cytometer (Acea-USA) and NovoExpress software (Acea-USA) for result analysis. The cell population was initially gated by FSC x SSC, following fluorescence analysis of EGFP using an FITC parameter (fluorescein isothiocyanate). Positive cell populations were gated and their fluorescence compared with parental cells as negative control. For each condition 10,000 events were acquired. The results were analyzed by measuring GFP intensity in each sample.

Statistics analysis

The statistical analysis and graphs were generated using the Graph Pad Prism 9.0 software. Results are expressed as the mean of triplicates +/− standard error of the mean. The groups were analyzed for variance using the one-way ANOVA test and post-tests by Turkey or Dunnett on experiment basis. Statistical analyses were considered significant, indicating p < 0.05.

Data availability

The data underlying this article are available upon reasonable request.

Acknowledgments

We thank LVV-LNBio for recombinant virus preparations and Dr. Clarice Weis Arns UNICAMP for providing ZIKV. We thanks Dr. Kleber Franchini for support and discussions. This research was funded by Fundação de Amparo à pesquisa do Estado de São Paulo (FAPESP), grants 2021/09107-9 and 2023/12245-0, Conselho Nacional de Desenvolvimento Cientifico e Tecnológico (CNPq) grant 440379/2016-4, fellowship 308403/2022-3, Brazilian Biosciences National Laboratory (LNBio), part of the Brazilian Center for Research in Energy and Materials (CNPEM), mainly funded by Brazilian Ministry for Science, Technology, and Innovations (MCTI). We also thank Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES) for PhD fellowship.

Author contributions

R.S.M.R. planned experiments, performed experiments, analyzed data, and wrote the manuscript. S.P. performed experiments and analyzed data. J.M.T. performed experiments. M.E.M. performed experiments. P.S.L.O. performed in silico experiments and discussed strategies and results. M.C.B. supervised the study, planned experiments, analyzed data, provided funding, and wrote the manuscript. All authors contributed to the article and approved the submitted version.

Declaration of interests

The authors declare no conflict of interest.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT in order to improve language and readability. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.