**4. Discussion**

Vector-borne diseases cause nearly one million deaths per year and represent 17% of all infectious illnesses worldwide [56]. This public health problem highlights the importance of understanding how arthropod vectors, microbes, and their mammalian hosts interact. At present, research efforts are focused on pathogen–host interactions, with a lack of attention on the significant contribution of vector-derived products in disease development. The molecular and cellular events occurring in vector–pathogen–host interactions are critical in determining the outcome of the vector-borne diseases. Recently, it was proposed that one strategy used by vectors, to promote a successful host infection, is the manipulation of EVs [41,57]. Infected vector cells secrete vesicles that may contain antigens, nucleic acids, and microbial cargos (or the whole pathogen), which exacerbate the pathogenesis and modulate the host responses [36]. EVs are involved in the exchange of bioactive molecules between cells. Although all EVs are vesicles constituted by lipid layers, their cargo reflects the state of the source cell, and their content can be altered in adverse conditions or manipulated by pathogens. Extracellular vesicles have an important role in the establishment of arboviral diseases [16,30,36]. EVs originating from arthropod vectors are an important strategy for immune evasion during viral transmission. For example, dengue virus (DENV) uses EVs derived from mosquitoes to infect mammalian cells. Mosquito-derived vesicles carry DENV proteins and a full-length viral genome. DENV transmission may occur through the interaction between the tetraspanin Tsp29Fb, a mosquito homolog of the human sEVs marker CD63, and the viral E protein [16]. Similarly, during the infection of the tick-cell line ISE6 with the Langat virus (LGTV), cells release sEVs that contain cargo from both the virus and the vector, which enable these EVs to transmit the virus to mammalian cells [36]. Ample evidence has been provided to show that sEVs carry and deliver viral genomes into recipient cells in vitro, as was reported for the hepatitis C virus (HCV), the hepatitis A virus (HAV), and human herpes virus 6 (HHV-6), among others [58–62]. Nevertheless, to date, it is unknown if, during mosquito ZIKV delivery to the vertebrate host cells, this arbovirus or the viral components such as viral proteins and the viral genome can be transferred by EVs from infected mosquito tissues to the host. We determined whether viral elements (viral RNA and envelope protein) could be transported by mosquito EVs, as was recently reported for other arthropod-borne flaviviruses.

In the present work, we found that, during ZIKV infection of C6/36 mosquito cells, small and large EVs were produced in high amounts. The isolated sEVs from ZIKV-infected cells were purified using paramagnetic nanobeads coated with anti-CD63 antibodies, thus demonstrating their endosomal origin. The separation of sEVs from virions by positive selection using magnetic beads coupled with an antibody against a tetraspanin enriched in sEVs is by far the best method [60,63]. The EVs characterization by nanoparticle tracking analysis (NTA) and transmission electron microscopy (TEM) from the mock C6/36 and the ZIKV-infected C6/36 EVs (Figures 2D and 3F) showed di fferent EVs populations, which were heterogeneous in shape, with sizes up to 200 nm (1000 nm scale) in the case of lEVs, and less than 200 nm (500 nm scale) in diameter for sEVs. These populations showed well-defined bilipid membranes in a proper structural resolution. Our data agree with other reports demonstrating that cells secrete EVs as a heterogeneous population with di fferent sizes and shapes [30,44,46]. Recently, it was proposed that the EVs size variation could be due to the internal content (proteins/RNAs) of vesicles produced from uninfected and infected cells [30]. By the NTA, we found that the lEVs from ZIKV-infected C6/36 presented an average size of 319.3 ± 11.5 nm (1.2-fold higher compared with the average size of lEVs from the mock cells) (Figure 2A). On the other hand, the sEVs from ZIKV-infected C6/36 presented an average size of 125.5 ± 1.6 nm (1.2-fold higher compared with the average size of sEVs from the mock cells) (Figure 3C). In this sense, the ZIKV C6/36 EVs RNase A + UV treatment assays (Figure 4 and Figure S4), performed to eliminate the possible presence of polluting free virus in samples, sugges<sup>t</sup> a presence of viral RNA inside ZIKV-infected C6/36 EVs. The presence of the ZIKV E protein detected in ZIKV C6/36 lEVs and sEVs isolates was not observed in mock C6/36 EVs (Figure 2E,F and Figure 3G,H). These findings sugges<sup>t</sup> that both sEVs and lEVs from ZIKV-infected mosquito cells carry the viral E protein and viral RNA. Different viruses manipulate EVs for their benefit in order to increase their persistence, pathogenesis, and transmission [58–62,64]. The hijacking of mosquito

cell membranes by ZIKV could facilitate their escape from host immune responses, promoting the viral elements' spread. We first used the epithelial cells (Vero) from the monkey *Cercopithecus aethiops* (used as gold standard cells for infection assay) [32]. The presence of lytic plaques (Figure 4E) in cultures of Vero with the ZIKV C6/36 EVs (lEVs and sEVs) isolates were observed in more concentrated amounts, formed in high or undetermined quantities. However, plaque formation was not detected in the negative controls. These data sugges<sup>t</sup> ZIKV-mosquito EVs participation during the infectious process of monkey epithelial cells. Additionally, the possibility of EVs participation during the infection process of naïve mosquito cells via cell-to-cell EVs transfer was evaluated. We demonstrated the presence of the ZIKV envelope protein at the mosquito cell membrane surface in cultures incubated with the ZIKV C6/36 EVs (large and small), implying the infection of naïve mosquito cells (Figure 5 and Figure S12). These results are consistent with those of Vora et al. (2018), showing that the full-length genome of DENV-2 detected in the EVs from DENV-infected mosquito cells was infectious in naïve mosquito and mammalian cells [16].

To date, it has not been determined whether EVs from ZIKV-infected mosquito cells are utilized as ZIKV viral element vehicles (genome/protein) to mammalian host cells. As monocytes and vascular ECs are the main targets during ZIKV human host infection, we evaluated the possible participation of EVs derived from the ZIKV-infected mosquito C6/36 cells in potential viral element carriers to human monocytes and vascular ECs. ZIKV was able to establish a productive infection in human monocytes and ECs, since the viral E protein was detected at high levels on the cell membrane surface at all PI time points of the assay (Figures 6 and 9). In stimulation assays of monocytes and ECs in the presence of ZIKV-infected C6/36 EVs, we found that ZIKV C6/36 EVs supported cell infection in naïve cells (Figures 7 and 10). Interestingly, the possible cellular mechanisms of the budding/tra fficking of sEVs that could be used by HCV, HAV, HIV, Epstein-Barr virus (EBV) and Kaposi's sarcoma-associated herpesvirus (KSHV) to spread from cell-to-cell viral elements or viral particles were recently revised [60]. Zhou et al. (2019), by using primary cultures of murine cortical neurons, showed that ZIKV uses sEVs as a mediator of viral transmission between neurons. Neuronal sEVs contained both ZIKV and viral RNA/protein(s) that were highly infectious to naïve cells. RNase A and neutralizing antibody assays suggested the presence of viral RNA/proteins inside EVs [65].

It is known that EVs may have di fferent functions when interacting with other cells, modifying their naïve cellular behavior [66]. For monocytes (Figure 8), in the presence of di fferent EVs samples, it was observed that naïve cells were di fferentiated and that they expressed a CD11b+ di fferentiation marker at high levels compared to the mock THP-1 cells and those stimulated with the mock C6/36 EVs. Likewise, the classical monocyte phenotype was changed to the CD14++ CD16+ intermediate phenotype with respect to the mock THP-1 cells and the mock C6/36 EVs stimulation assays (*p* < 0.0001 for CD14 and *p* < 0.01 for CD16). Intermediate and non-classical monocytes seem to be the main producers of pro-inflammatory mediators in response to viral infection [17,49]. The activation of naïve monocytes via the ZIKV C6/36 sEVs favors TNFα mRNA expression, which suggests that ZIKV C6/36 EVs–human monocyte interplay plays a role in establishing a pro-inflammatory state.

The possible participation of ZIKV C6/36 EVs in the infection and activation of vascular ECs was also evaluated. In cultures performed in HMEC-1 naïve cells in the presence of ZIKV (MOI 1), the mock C6/36 EVs, and the ZIKV small/large C6/36 EVs, the amount of ZIKV E protein was evaluated. Figure 10 shows the viral E protein presence at high levels in the ZIKV (MOI 1)-infected HMEC-1 cells as well as in the naïve ECs stimulated with ZIKV C6/36 EVs (RNase A + UV-treated and untreated). The E protein percentages were compared among values of all conditions against the mock HMEC-1 cells; these values were statistically significant (*p* < 0.0001) (Figure 10B,C). These data sugges<sup>t</sup> that ZIKV-infected C6/36 EVs support the infection of mammalian host cells, including vascular ECs.

At the endothelial vascular cell level, our data indicate that ECs are susceptible to ZIKV infection and activation by ZIKV C6/36 EVs, and these EVs favor the induction of damage receptors, such as coagulation (TF) and inflammation (PAR-1) receptors, and adhesion molecule (ICAM-1) presence at the cell membrane's surface level (Figure 11). Recent observations in human and animal models [51,52,67] sugges<sup>t</sup> that, in severe Zika cases, di fferent coagulation disorders occur. It has been shown that several

viruses activate the coagulation system especially through TF receptor expression [53]. Dengue virus has also been shown to cause coagulation disorders in ECs [54].

Anfasa et al. (2019) provided in vitro evidence that ZIKV infection of human umbilical vein endothelial cells (HUVECs) induces apoptosis and increases TF production, which triggers the activation of secondary hemostasis [68]. Additionally, to evaluate the possible expression of a pro-inflammatory response by EVs released from ZIKV-infected C6/36 cells, we measured the TNFα mRNA expression in naïve ECs infected by ZIKV (MOI 1) and those stimulated with the mock C6/36 EVs or the ZIKV C6/36 EVs (small/large isolates the same for RNase A + UV-treated). We found that, like ZIKV infection, the ZIKV C6/36 EVs were able to induce TNFα mRNA expression in endothelial vascular cells (Figure 11D). The ZIKV C6/36 EVs also participated in vascular endothelial damage, with a weakening of the endothelial barrier integrity, and this was demonstrated using a Transwell assay (Figure 11E). We previously reported that DENV infection of ECs upregulates the TF coagulation receptor in endothelial vascular cells, which triggers the generation of hemostatic proteases (thrombin) favoring the activation of protease-activated receptors or PARs, which, in turn, induces signaling inflammatory pathways (via phosphorylation of MAPKs p38 and ERK1/2, by transcription of the NF-κB factor), thereby supporting the upregulation of adhesion VCAM-1 or pro-inflammatory molecules in ECs, being part of the pathogenic mechanisms for the vascular endothelial injury present in severe Dengue cases [54]. At present, it is under evaluation whether these signaling pathways can support the participation of ZIKV-infected vector cells in the activation and damage of vascular ECs that can contribute to the pathogenesis of severe ZIKV cases. The present data sugges<sup>t</sup> that ZIKV C6/36 EVs allow for ZIKV elements (viral genome/protein) to modulate host response and enhance viral fitness abilities.

In summary, during ZIKV infection of C6/36 mosquito cells, small and large EVs were produced (Figure 12A). The mosquito EVs released from ZIKV-infected cells carried viral E protein and viral RNA and were able to infect naïve mosquito and mammalian cells. The ZIKV-infected mosquito EVs, then, modified the naïve cellular behavior, since they promote the infection, activation, and differentiation of human monocytes that favor a pro-inflammatory monocyte state (Figure 12B). At the endothelial vascular cell level, our data indicate that vascular ECs are susceptible to ZIKV activation and infection by C6/36 EVs, which favor the induction of tissue damage receptors, such as coagulation (TF) and inflammation (PAR-1) receptors, and adhesion molecule presence (ICAM-1) at the membrane surface level with an increase in cell permeability (Figure 12C). Knowledge of the targets' cellular pathways that allow ZIKV to establish prolonged viral persistence could contribute to novel vaccines and therapies.

**Figure 12.** Extracellular vesicles (EVs) from ZIKV-infected mosquito (C6/36) cells participate in the modification of naïve cells' behavior by mediating cell-to-cell transmission of viral elements (Graphic description). ( **A**) ZIKV C6/36 EVs favor naïve mosquito cell infection. Mosquito image created with BioRender.com. (**B**) ZIKV C6/36 EVs promote infection and shift to a pro-inflammatory phenotype in monocytes. ( **C**) ZIKV C6/36 EVs participate in the vascular EC infection and activation.

**Supplementary Materials:** The supplementary materials are available online at http://www.mdpi.com/2073-4409/9/1/ 123/s1.

**Author Contributions:** Conceptualization: P.P.M.-R. and B.H.R.-O.; methodology: P.P.M.-R., E.Q.-G., V.M.-M., and B.H.R.-O.; validation: P.P.M.-R., E.Q.-G., and V.M.-M.; formal analysis: P.P.M.-R. and B.H.R.-O.; investigation: B.H.R.-O.; resources: B.H.R.-O, L.T.A.-M., and L.F.J.-G.; data curation: L.T.A.-M. and L.F.J.-G.; writing—original draft preparation: P.P.M.-R. and B.H.R.-O.; writing—review and editing: MDPI; visualization: P.P.M.-R. and B.H.R.-O.; supervision: B.H.R.-O.; project administration: V.M.-M.; funding acquisition: B.H.R.-O. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Acknowledgments:** Pedro Pablo Martínez-Rojas is a doctoral student from Programa de Doctorado en Ciencias Bioquímicas, Universidad Nacional Autónoma de México (UNAM), and was supported by Consejo Nacional de Ciencia y Tecnología (CONACYT), México. Elizabeth Quiroz-García is a doctoral student from Programa de Doctorado en Ciencias Biomédicas, UNAM and was supported by CONACYT. We thank the PAPIIT-UNAM Program for partially funding support. We also thank LabNalCit-UNAM (CONACYT) for the technical support in the acquisition of flow cytometry samples and Miguel Tapia Rodríguez, from Unidad de Microscopía, Instituto de Investigaciones Biomédicas UNAM, for his support in the use of the Olympus IX71 microscope.

**Conflicts of Interest:** The authors declare that there is no conflict of interest.
