U87-MG is glioblastoma cells, kindly provided by Dr. Olivia Hernández-González (Instituto Nacional de Rehabilitación [INR], Mexico), were cultured in Advanced DMEM (GIBCO, 12491015) supplemented with 13% fetal bovine serum (FBS, GIBCO, 12657-02), 1 U/mL penicillin/streptomycin (Sigma, P4333), and 2.5 mM glutamine (Sigma-Aldrich, G8540). Experiments with U87-MG cells were performed in three independent replicates.

All experimental animal procedures followed the Mexican Official Standard for the Production, Care, and Use of Laboratory Animals (NOM-062-ZOO-1999). The protocols for the use of AG129 mice (129/Sv strain; 6–8 weeks old, both male and female), obtained from Marshall Bioresources, UK (protocol 048-02), and neonatal CD1 mice (5–7 days old) (ICR-CD1, strain code: 022) (protocol 0382-24) were approved by the Institutional Animal Care and Use Committee (IACUC) of CINVESTAV-IPN.

AG129 and CD1 mice were housed under controlled conditions in the Animal Production and Experimentation Unit (UPEAL-CINVESTAV) with ad libitum access to food and water. Euthanasia of adult mice and neonates was performed in a 10-liter CO₂ chamber, using a displacement rate of 30% to 70% of the chamber volume per minute, in accordance with the euthanasia of rodents using carbon dioxide following AVMA (American Veterinary Medical Association) guidelines. For adult mice, CO₂ was administered for 5 minutes; for neonates, the exposure time was extended to 10 minutes due to their greater resistance to hypoxia. In both cases, a physical method of decapitation was applied after CO₂ exposure to ensure death. The entire procedure was conducted under veterinary supervision and in compliance with institutional protocols and current ethical regulations.

Zika virus (ZIKV, MEX_CIENI551 strain) was kindly provided by Dr. Jesús Torres (Escuela Nacional de Ciencias Biológicas [ENCB], Instituto Politécnico Nacional, Mexico), and dengue virus serotype 4 (DENV-4, H241 strain) was donated by the Dr. Manuel Martínez Báez Institute of Epidemiological Diagnosis and Reference (InDRE, Mexico).

Both viruses were propagated in the brains of neonatal CD1 mice at the UPEAL-CINVESTAV facility. Lysates from infected and uninfected tissues were prepared and used to infect U87-MG cells at a multiplicity of infection (MOI) of 5. In vivo experiments were conducted by inoculating AG129 and CD1 mice with 1 × 10⁶ focus-forming units (FFU) of ZIKV or 1 × 10⁷ FFU of DENV-4. Negative control groups were administered an equivalent volume of saline solution under the same experimental conditions.

Cells were treated with melatonin (Sigma-Aldrich, M5250), which was dissolved in 40% ethanol at a concentration of 125 mM and stored in aliquots at −20°C until use. Final concentrations of 0.6 mM and 1.2 mM were initially tested in vitro; subsequently, the 1.2 mM dose was selected for all further experiments. To inhibit the RhoA pathway, Rho Inhibitor I (Cytoskeleton, Clostridium difficile toxin B [CT04]) was reconstituted according to the manufacturer’s instructions, aliquoted, and stored at −20°C. The final working concentration was 1.0 μg/mL. In in vivo experiments, liquid melatonin (Vitamatic, USA) was administered at 20 mg/kg to AG129 mice and 10 mg/kg to neonatal CD1 mice, according to body weight, without altering the original formulation.

Cell viability in the presence of melatonin (MEL) was evaluated in U87-MG cells using the colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Sigma-Aldrich). Cells were seeded in flat-bottom 96-well plates and treated with various concentrations of MEL (0.15, 0.3, 0.6, and 1.2 mM) for 24 and 48 hours. Following treatment, MTT reagent was added to each well, and plates were incubated at 37°C in a 5% CO₂ atmosphere for the time specified by the manufacturer’s instructions. Formazan crystals were then solubilized with DMSO, and absorbance was measured at 562 nm, with 630 nm as the reference wavelength to correct for background absorbance.

Vehicle-treated cells (40% ethanol) were used as controls and were considered to represent 100% viability. Across all tested concentrations, cell viability remained above 80% after both 24 and 48 hours of treatment.

U87-MG cells were cultured on coverslips in 24-well plates until reaching 70–80% confluence. Cells were infected or mock-infected with ZIKV or DENV-4 by exposure to the virus for 2 hours. Following infection, cells were washed, replaced with fresh medium and treated with MEL (0.6 or 1.2 mM) for 48 hours post-infection (hpi).

For experiments involving the RhoA inhibitor CT04, U87-MG cells were cultured under the same conditions, infected as described, and then incubated with serum-free DMEM supplemented with 1.0 μg/mL CT04 for 48 hpi.

At the end of the treatments, cells were washed three times with 1× PBS, fixed with 4% paraformaldehyde (PFA) for 30 minutes at 4 °C, and permeabilized with 0.2% saponin, 1% fetal bovine serum (FBS), and 1× PBS for 30 minutes at room temperature. Cells were then incubated overnight at 4 °C with primary antibodies specific for viral infection detection: ZIKV anti-NS3 protein (1:300, Genetex, GTX133309), DENV anti-NS3 protein (1:300, Genetex, GTX124252) and anti-IBA1 (1:200, Santa Cruz, sc-32725), Alexa Fluor 555-conjugated anti-rabbit secondary antibody (1:300, Life Technologies) and Alexa Fluor 488-conjugated anti-rabbit secondary antibody (1:750, Life Technologies), were used to detect the primary antibodies, and nuclei were counterstained with Hoechst (1:1000, Santa Cruz Biotechnology). Images were acquired using a Leica TCS SP8 confocal microscope (Leica Microsystems) and processed with Leica Application Suite X Core Offline v3.3.0 software.

For brain and cerebellum sections of AG129 and CD1 mice, frozen tissue from the left lobe of three mice per condition was processed according to the protocol described by Trist, Wright, and Rangel (12). The following antibodies were used to detect viral infection and cellular markers of inflammation: ZIKV anti-NS3 protein (1:300, Genetex, GTX133309), DENV anti-NS3 protein (1:300, Genetex, GTX124252), anti-Iba1 (1:200, Santa Cruz, sc-32725), anti-CD86 (1:150, Cell Signaling, 19589) and anti-CD206 (1:150, Cell Signaling, 24595). An Alexa Fluor 555-conjugated anti-rabbit secondary antibody (1:300, Life Technologies) and Alexa Fluor 488-conjugated anti-rabbit secondary antibody (1:750, Life Technologies), were used to detect the primary antibodies, and nuclei were counterstained with Hoechst (1:1000, Santa Cruz Biotechnology).

Samples were analyzed using the Leica SP8 confocal microscope. Image processing was performed using LASX Office 1.4.7–28921 software. Three fields from three independent experiments were analyzed.

A focus-forming unit (FFU) assay was performed to quantify the number of infectious viral particles produced in MEL-treated cells. Supernatants from infected and mock-infected cells were processed according to. The FFU assays were performed to quantify infectious ZIKV and DENV-4 particles. Vero cells were seeded in 96-well plates and infected with serial dilutions of viral supernatants for 2 h at 37°C. After infection, cells were overlaid with a medium DMEM and incubated for 48 h. Cells were then fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100. Viral foci were detected by staining with the 4G2 antibody (anti- Envelope 1:300, Genetex, GTX57154) and analyzed by epifluorescence microscopy using a Leica DM IL microscope.

In vitro assays were performed to quantify viral genome copy number and evaluate the expression of proinflammatory and interferon-stimulated genes (ISGs) from cell monolayers. Total RNA was extracted using TRIzol Reagent (Invitrogen, Cat. 15596026.PPS). For in vivo analysis, from 50 mg of tissue collected from the right cerebral lobe, including the cerebellum, of three mice per condition (each mouse considered an individual experimental unit). These samples were used to determine viral load and assess the expression of immune response-related genes. According to the manufacturer’s instructions. Genomic DNA contamination was eliminated by treatment with DNase I (New England Biolabs, Cat. M0303).

Quantitative real-time PCR (qPCR) was performed to quantify ZIKV and DENV-4 RNA copies in brain and cerebellar tissues. Reactions were carried out in a final volume of 10 μL containing 5 μL of qPCRBIO SyGreen BlueMix Lo-ROX (PCR Biosystems, PB20.15), 0.4 μL of each primer (10 μM) ( Table 1 ), and 1 μL of diluted cDNA, with RNase-free water added to adjust the volume.

Thermocycling conditions were: 50°C for 10 min, 95°C for 2 min, followed by 40 cycles of 95°C for 30 s and 60°C for 30 s.

The assay was performed on an Eco Real-Time PCR System (Illumina), and quantification was achieved using a standard curve generated from serial dilutions (10¹⁰ to 10² copies) of a 220-bp fragment of the ZIKV and DENV-4 genomes, amplified using general Orthoflavivirus primers as described by Maher-Sturgess et al. (2008). The detection threshold was established using RNA from an uninfected control mouse. Data were analyzed with EcoStudy software (v5.048909) and expressed as viral copies/50 mg of tissue.

RhoA pathway activation was assessed using the RhoA Activation Assay Kit (Cytoskeleton, Denver, CO) according to the manufacturer’s instructions.

U87-MG cells were cultured in 24-well plates until reaching 80% confluence. Cells were then infected with ZIKV or DENV-4 at a multiplicity of infection (MOI) of 5 and treated with 1.2 mM MEL for 18, 24, or 48 hpi. Control groups included uninfected cells and cells treated with HBSS medium alone. Three independent experiments were conducted for each condition.

At the indicated time points, cells were lysed and immediately frozen in liquid nitrogen, then stored at −80°C until further processing.

AG129 mice are widely used as a canonical model for ZIKV and DENV infections, primarily due to their high susceptibility to flavivirus infection and their capacity to exhibit consistent and observable clinical signs of disease. The mice were randomly selected for each experimental group, separated only by sex. AG129 mice were infected intraperitoneally with 4×10⁶ PFU/mL of ZIKV or 4×10⁷ PFU/mL of DENV-4, delivering 100 μL of inoculum per mouse. Control mice were inoculated with virus-free tissue lysates.

MEL treatment started four days post-infection (dpi). Mice in the treatment groups received 20 mg/kg of MEL daily via oral gavage, while infected control mice received the same volume of water. Methodology was adapted and modified from Cecon et al. (8). Body weight and clinical signs of disease were recorded daily until euthanasia. Disease severity was scored using a clinical scoring system described by Orozco et al. (13), where 1 indicates healthy mice and 5 represents moribund animals ( Supplementary Figure 1 ).

Three independent experiments were performed, each including one or two mice per group. Animals were excluded from the analysis if death occurred before the expected onset of clinical signs associated with viral infection (i.e., before day 3 post infection), in the absence of weight loss or neurological symptoms, and without detectable viral RNA in serum or brain tissue (assessed by qRT-PCR). These criteria allowed us to distinguish deaths likely due to procedural problems (e.g., complications related to gastric tube or stress) from those attributable to the infection Specifically, animals that reached a cumulative clinical score of 5—based on weight loss, neurological signs, and general condition—were humanely euthanized to prevent unnecessary suffering. Euthanasia was performed using a CO₂ chamber.

Survival data, clinical scores, and body weight were graphically represented and analyzed using GraphPad Prism 10 software.

CD1 mice aged 5–7 days were infected intraperitoneally with 1×10⁷ PFU/mL of ZIKV or DENV-4, diluted in physiological saline to a final volume of 20 μL per mouse.

The experimental design included a time-course analysis, with day 0 corresponding to the day of inoculation. MEL treatment (10 mg/kg) was administered daily by oral gavage from day 1 to day 7 post-infection (dpi), following the protocol described by Cecon et al. and Francis et al. (8, 14). Mice were euthanized and collected at days 3 (n=9), 5 (n=9), and 7 (n=9) dpi.

Two independent experiments were performed using litters of three mice per virus at each time point. Viral load and the expression of immune response-related genes were assessed by qPCR as described above. Additionally, frozen brain sections were processed and immunolabeled for analysis by confocal microscopy. Data were graphically represented and analyzed using GraphPad Prism 10 software.

Confocal microscopy images were analyzed using LASX Office software (v1.4.7 28921). One-way ANOVA followed by Mann–Whitney test was used to compare group means in the FFU assays, qPCR results, RhoA activity assays, and mean fluorescence intensity (MFI) measurements.

For in vivo assays, Kaplan-Meier survival curves were generated, and survival rates between treated and untreated groups were compared using the Wilcoxon and Mantel-Cox tests. The ANOVA-LSD test was also used to compare mean survival times between groups. A p-value ≤ 0.05 was considered statistically significant in all analyses.

The RhoA signaling pathway is critical for neurodevelopment and regulates various cellular processes, including cytoskeletal remodeling and immune responses. Previous studies have demonstrated that several viruses, including EBOV, respiratory syncytial virus (RSV), rotavirus (RV), porcine pseudorabies virus (PRV), and SARS-CoV-2, hijack RhoA pathway to facilitate infection (10, 15–17). The concentration of 1 μg/mL was recommended by the manufacturer for use in neural-derived cells such as U87-MG. Additionally, a cytotoxicity analysis using the MTT assay confirmed that this concentration is safe for use in our cell cultures ( Supplementary Figure 2 ). In many cases, RhoA activation is proviral, whereas its inhibition leads to antiviral effects. Our results demonstrate a direct relationship between RhoA activation and the replication of ZIKV and DENV-4. Immunofluorescence microscopy revealed that a dose of 1 μg/mL led to a reduction in viral protein expression (ZIKV, green; DENV-4, red) upon RhoA inhibition ( Figures 1A, F ), accompanied by decreased mean fluorescence intensity (MFI) ( Figures 1B, G ). Reductions in infectious viral particles, as measured by focus-forming units (FFU/mL) ( Figures 1C, H ), and viral genome copy numbers by qPCR ( Figures 1D, I ) further supported these findings. Furthermore, a reduction in the active form of RhoA was confirmed at 48 hpi following treatment with 1 µg/mL of CT04 ( Figures 1E, J ).

These results suggest that CT04, a RhoA pathway inhibitor, interferes with the replication of ZIKV and DENV-4.

MEL is a hormone produced by the pineal gland, widely used for managing sleep disorders, anxiety, and seasonal depression (4, 18). Recent studies have revealed its antioxidant, anti-inflammatory, and immunomodulatory properties (8, 9), suggesting potential antiviral effects. To explore its antiviral potential, U87-MG cells were treated with increasing concentrations of MEL (0.15, 0.3, 0.6, and 1.2 mM) and subjected to cytotoxicity assays (MTT) at 24 and 48 hours post-treatment. In all conditions, cell viability remained above 80% ( Supplementary Figure 3 ), indicating that the tested concentrations were safe for further analysis. Based on these results, 0.6 and 1.2 mM were selected for subsequent experiments.

U87-MG cells infected with ZIKV or DENV-4 (MOI 5) and treated with 0.6 or 1.2 mM MEL, showed a marked reduction in viral protein expression by confocal microscopy ( Figures 2A–E ), as evidenced by decreased MFI ( Figures 2B, F ). FFU assays ( Figures 2C, G ) and qPCR analyses ( Figures 2D, H ) confirmed significant reductions in viral titers and genome copy numbers, respectively. These findings indicate that MEL inhibits ZIKV and DENV-4 replication in U87-MG cells, with a stronger effect at 1.2 mM.

Given the role of ZIKV and DENV-4 in glial infection and neuroinflammation, we assessed the impact of MEL on inflammatory gene expression in infected U87-MG cells. Treatment with 1.2 mM MEL significantly reduced the expression of key pro-inflammatory genes (IL-1β and TNFα) and ISGs (MX1, IFI44L, and IFNβ) at 48 hpi. During ZIKV infection ( Figures 3A–E ), MEL reduces expression of TNFα, MX1 and IFI44L but significantly increases expression of IL-1β and IFNβ. In contrast, during DENV-4 infection ( Figures 3F–J ), MEL consistently reduced the expression of all evaluated genes. These results suggest that melatonin exerts virus-specific immunomodulatory effects during Orthoflavivirus infections.

Considering that both ZIKV and DENV-4 manipulate cytoskeletal dynamics during infection, we evaluated whether MEL modulates RhoA activation. Confocal microscopy revealed that a 1.2 mM dose effectively reduced viral protein expression (ZIKV envelope and DENV-4 NS3) ( Figures 4A, B ), with a corresponding decrease in MFI at 48 hpi ( Figure 4C ). qPCR analysis corroborated a sustained decrease in viral genome copies over time (18, 24, and 48 hpi) ( Figure 4D ).

Kinetic ELISA analysis demonstrated significant activation of RhoA upon ZIKV and DENV-4 infection, which was significantly reduced by melatonin treatment at 24 and 48 hpi ( Figure 4E ). These findings suggest that melatonin inhibits viral replication at least partly through suppression of RhoA signaling.

To evaluate the protective effect of MEL in an in vivo model, AG129 mice were infected with ZIKV or DENV-4 and treated with 20 mg/kg MEL. In ZIKV-infected mice ( Figures 5A–F ), mild symptoms appeared early, progressing rapidly to severe neurological impairment and death. Although no significant differences in survival were detected ( Table 2 ), melatonin-treated mice exhibited delayed symptom onset and slightly. In the DENV4 model, MEL treatment significantly increased the mean survival from 12.25 to 16.25 dpi and Along with delayed clinical signs and maintained body weight and lower viral genome copies in the brain. In the ZIKV model, although survival differences were not statistically significant, MEL-treated mice showed a slight increase in average survival 7.6 to 6.75 dpi ( Table 2 ), along with delayed clinical signs, and to body weight maintenance ( Figures 5B, C ). Moreover, quantitative analysis revealed lower viral loads of ZIKV and DENV-4 in the brains of animals treated with melatonin.

In DENV-4 infected mice ( Figures 5A–F ), melatonin significantly improved survival rate (16.25 vs. 12.25 days, treated and untreated, respectively; Table 2 ), delayed symptom onset, preserved body weight, and reduced brain viral loads ( Figure 5G ). Fluorescent immunohistochemistry (IHC) of brain and cerebellar tissues from DENV-4-infected male and female mice revealed greater IBA1 activation in females compared to males, correlating with observed survival outcomes ( Figure 5H ). MEL-treated mice exhibited decreased IBA1 and NS3 signals, suggesting reduced neuroinflammation and viral replication. The slower progression of disease in DENV-4 infection may have contributed to the enhanced therapeutic efficacy of melatonin. In ZIKV-infected mice ( Figure 5H ), IBA1 expression was higher in males compared to melatonin-treated counterparts, while females displayed stronger NS3 signals, indicative of higher viral burden. Importantly, both male and female mice treated with melatonin showed reduced IBA1 and NS3 expression, supporting its potential role in limiting neuroinflammation and viral replication in the brain and cerebellum.

Finally, we tested melatonin in a transient infection model using neonatal immunocompetent CD1 mice. Following intraperitoneal infection with ZIKV or DENV-4, viral RNA was detectable at 3 dpi and peaked at 7 dpi (10⁷ copies for ZIKV; 10⁸ for DENV-4) ( Figure 6A ). MEL treatment (10 mg/kg) substantially reduced viral loads by 3–4 log units ( Figure 6A ).

Expression analysis of neuroinflammatory markers revealed that melatonin reduced IL-1β, TNFα, and NF-κB during ZIKV infection ( Figures 6B–D ). During DENV-4 infection, melatonin reduced TNFα but increased IL-1β and NF-κB expression at 7 dpi ( Figures 6E–G ). These findings suggest that MEL can modulate neuroinflammation and inhibit viral replication in the cerebral cortex, supporting its therapeutic potential against Orthoflavivirus induced neuroinfections.

Several studies have shown that the RhoA signaling pathway is involved in regulating M1 and M2 polarization of macrophages and microglia (19–21). Although the precise mechanisms underlying this phenotypic switch and cellular activation remain debated, few investigations have directly addressed the role of RhoA as a neuroprotective factor during ZIKV and DENV-4 neuroinfections.

To further explore this, we assessed proinflammatory and anti-inflammatory markers in cerebral cortex of neonatal immunocompetent CD1 mice at 7 dpi. IBA1 (green) and CD86 (purple) were used to identify the proinflammatory M1 microglial phenotype, while CD206 (yellow) served as a marker of the anti-inflammatory M2 phenotype. Our results revealed distinct neuroimmune responses between ZIKV and DENV-4 infections. ZIKV infection was associated with increased expression of IBA1 and CD86, indicative of predominant M1 microglial activation. However, melatonin treatment shifted this profile toward an M2 phenotype, as evidenced by elevated CD206 expression, suggesting that melatonin promotes inflammation resolution and tissue repair during ZIKV infection ( Figure 6H , right).

In contrast, DENV-4 infection induced a mixed microglial response, characterized by increased IBA1 and CD206 expression and low CD86 levels. Notably, melatonin treatment in DENV-4 infected mice resulted in decreased IBA1 and CD206 expression, alongside increased CD86 levels, indicating a shift toward a proinflammatory M1 phenotype ( Figure 6H , left). These findings support the notion that melatonin exerts differential immunomodulatory effects in the brain, promoting an M2 phenotype in ZIKV infection and an M1 phenotype in DENV-4 infection. This differential modulation aligns with our qPCR data, which showed increased expression of proinflammatory genes such as IL-1β and NF-κB following melatonin treatment ( Figures 6B–G ).