*3.1. ZIKV Infects C6*/*36 Mosquito Cells*

The ZIKV infection of C6/36 mosquito cells was evaluated by the presence of the ZIKV E protein (FACS) at the cell membrane surface at 24, 48, 72, 96, and 120 h PI. We found that, when using a multiplicity of infection (MOI) of 1 at 48 h PI, the mosquito cells had high levels (38.60 ± 1.05%) of ZIKV E protein at the membrane surface level (Figure 1A).

Likewise, a high percentage (40%) of ZIKV-infected C6/36 cells at 48 h PI was present (Figure 1B), when comparing the E protein levels with the mock cell, and these data were statistically significant (*p* < 0.0001). After 48 h PI, less fluorescent cells were observed in the microscopy assay (Figure S7). It could be that, after main viral infection time, most syncytia structures were in development. The ZIKV-infected C6/36 cultures (48 h PI) developed a more cytopathic effect with the formation of syncytium structures but without the detachment of the cell monolayer, showing increased viral E protein-positive cells determined by fluorescence microscopy (Figure S7). Since the optimal cell infections (activation) were obtained in these conditions (MOI 1, 48 h PI), they were used for all subsequent infection experiments. Recently, it was shown that, during the virus infection process, infected cells are activated

and produce different subtypes of EVs, which may have different functions when interacting with other cells, modifying naïve cellular behavior [41]. Therefore, we evaluated whether ZIKV-infected C6/36 mosquito cells could release large and/or small EVs. EVs produced from other arbovirus-infected cells were able to mediate the cell-to-cell communication between vector-host cells [36].

**Figure 1.** Zika virus (ZIKV) (multiplicity of infection (MOI) 1) infects C6/36 cells. (**A**) ZIKV envelope (E) protein detection at 24, 48, 72, 96, and 120 h post-infection (PI) by FACS assay. Dot plots are the representative mean ± standard deviation (SD) of the positive cells from three independent experiments. (**B**) ZIKV-infected cells percentages obtained by FACS. The ZIKV E protein levels were compared (by an unpaired Student's t-test) with the mock C6/36 (\*) value. Statistical significance was recognized as \*\* when *p* < 0.01, and \*\*\* when *p* < 0.0001.

#### *3.2. ZIKV-Infected C6*/*36 Cells Release Large EV Phosphatidylserine*+ *ZIKV E Protein*+

Large EVs were developed by the outward shedding of the cell plasma membrane during cell activation and have a size greater than 200 nm, which can be identified by an Annexin-V binding assay (by FACS). This assay exposes phosphatidylserine (PS) on the outer plasma membrane leaflet [42,43]. The lEVs released from mosquito ZIKV-infected cells and mock cells were isolated from culture supernatants as described above. The characterization of all large EVs was performed by nanoparticle tracking analysis (NTA) by using the NanoSight NS300 equipment and Malvern Instruments software. The different experimental conditions for EVs detection were first established (Figure S2A), and quantitative controls with 100 and 200 nm polystyrene microspheres (NTA4088 and NTA4089, Malvern) were used (Figure S2B). Nanoparticles present in PBS and EVs-depleted FBS were also quantified to correct the number of the isolated EVs. The ZIKV virions were detected (as a peak of size of 63.5 ± 8.1 nm) to identify its presence in the EVs isolates from ZIKV-infected mosquito cells (Figure S2C).

For the NanoSight 300, a camera level of 12.0 was used with a detection limit of 2.5, a temperature of 20 ◦C, samples diluted in 1.0 mL of PBS, reading periods of 30 s, and three consecutive repetitions. The NTA from lEVs isolates of ZIKV-infected cells showed a concentration of 2.92 × 10<sup>10</sup> ± 3.55 × 10<sup>9</sup> particles/mL with an average value of size of 319.3 ± 11.5 nm. Compared with NTA from lEVs isolates of mock cells (2.14 × 10<sup>9</sup> ± 4.10 × 10<sup>8</sup> particles/mL with an average value of size of 268.9 ± 8.2 nm), the concentration of nanoparticles from lEVs isolates from infected cells were 13.6-fold higher (Figure 2A).

In parallel, we observed that, during ZIKV infection (MOI 1, 72h PI) of the C6/36 cells (Figure 2B), the percentage of lEVs PS+ released to the cell culture supernatant (60.50 ± 0.87%) was 1.98-fold higher (*p* < 0.001) than lEVs PS+ from uninfected (30.60 ± 0.96%) mosquito cells (Figure 2C). The fluorescence emitted by Annexin-V binding was proportional to the PS presence: The ZIKV-infected C6/36 lEVs mean fluorescence intensity (MFI) was compared with that of the microbead control and mock C6/36 lEVs values (Figure S8A). The Annexin-V binding MFI in the lEVs from ZIKV-infected cultures was 1.45-fold higher than that of the lEVs from the mock cultures.

**Figure 2.** ZIKV-infected C6/36 cells issue the large extracellular vesicles (lEVs) phosphatidylserine (PS)<sup>+</sup>. (**A**) Nanoparticles tracking analysis (NTA) of the purified lEVs isolates from the mock and ZIKV-infected C6/36 cells. Histograms are the representative mean ± SD of the nanoparticle's concentration (particles/mL) and the size (nm) from three independent experiments. (**B**) PS detection by the Annexin-V binding assay. The fluorescence emitted by Annexin-V binding is proportional to PS levels. FACS dot plots are the representative mean ± SD of the lEVs PS+ from three independent experiments. (**C**) lEVs PS+ percentages obtained by FACS. The PS levels were compared (by an unpaired Student's t-test) with the mock lEVs (\*) value. (**D**) Transmission electron microscopy (TEM) images from the mock C6/36 and the ZIKV-infected C6/36 lEVs (1000 nm scale). (**E**) ZIKV E protein detection on the lEVs surface by FACS assay. Dot plots are the representative mean ± SD of the lEVs ZIKV E protein+ from three independent experiments. (**F**) Percentages of the lEVs ZIKV E protein+ by FACS. The ZIKV E protein levels were compared (by an unpaired Student's t-test) with the mock C6/36 (\*) value. Statistical significance was recognized as \*\* when *p* < 0.01, and \*\*\* when *p* < 0.0001.

The lEVs samples were also characterized by transmission electron microscopy (TEM): The analysis of images from the mock C6/36 lEVs and ZIKV-infected C6/36 lEVs (Figure 2D) showed different EV populations, which were heterogeneous in shape, with sizes up to 200 nm (1000 nm scale) and a defined

membrane in a proper structural resolution. These data agree with other reports, which demonstrated that cells secrete EVs as a heterogeneous population with different sizes and shapes [30,44]. We did not identify viral particles inside lEVs TEMs from ZIKV-infected cells, so we evaluated the ZIKV E protein presence on the membrane's surface of lEVs. We found 18.27 ± 1.27% of positive ZIKV E protein+ lEVs (*p* < 0.01) compared with the lEVs from the mock C6/36 cells (Figure 2E,F). The ZIKV E protein (MFI) in the lEVs from ZIKV-infected cells was 18.6-fold higher than that of the lEVs from the mock cultures (Figure S8B).

#### *3.3. ZIKV-Infected C6*/*36 Cells Release Small EV CD63-Like*+ *ZIKV E Protein*+

First, the presence of CD63-like tetraspanin at the membrane surface of C6/36 cells was evaluated as well. The point time of the CD63 membrane decreased (CD63 internalization) to select the best condition for the identification of the sEVs marker (inside cells) and, likewise, the time for the optimal isolation of sEVs CD63-like+ [45]. The tetraspanin detection (FACS) was evaluated at 24, 48, 72, 96, and 120 h PI (Figure 3A). We observed that naïve C6/36 cells constitutively contain high levels of the CD63-like tetraspanin (Figure 3A,B) at the cell membrane surface (55.67 ± 2.30%). Nevertheless, in ZIKV-infected cultures, an increase in the CD63 percentage (79.40 ± 2.78%) was present at 24 h PI (1.4-fold higher), while the highest percentage (88.73 ± 1.35%) of the CD63+ cells were obtained at 48 h PI (1.6-fold higher compared with mock cells). The increased tetraspanin values were significant (*p* < 0.0001). We found that CD63 levels were decreased at 72–96 h PI (Figure 3A,B). These data sugges<sup>t</sup> that CD63-tetraspanin internalization may occur after 48 h PI. Therefore, the sEVs biogenesis in mosquito cells could take place between 48 and 72 h. The sEVs CD63+ were then isolated at 72 h PI. The presence of the CD63 tetraspanin was also evaluated (red) by fluorescence microscopy (100×) in mock and ZIKV-infected C6/36 cultures (green for ZIKV E protein). The presence of the CD63 tetraspanin inside C6/36 cells suggests the endosomal nature of small EVs (Figure S9A).

The NTA from sEVs isolates of ZIKV-infected cells showed a concentration of 3.17 × 10<sup>11</sup> ± 5.62 × 10<sup>10</sup> particles/mL with an average value of size of 125.5 ± 1.6 nm. Compared with the NTA from sEVs isolates of mock cells (2.39 × 10<sup>10</sup> ± 4.41 × 10<sup>9</sup> particles/mL with an average value of size of 107.8 ± 3.1 nm), the concentration of nanoparticles from sEVs isolates from infected cells were 13.3-fold higher (Figure 3C).

The sEVs isolates were then identified by positive selection, using paramagnetic nanobeads coated with anti-CD63 antibodies (Figure S3). Previously, we determined whether ZIKV viral particles could cross-react with the paramagnetic nanobeads (Figure S9B,C). We found that ZIKV did not couple to the nanobeads, so the sEVs CD63+ detection is free of viral particles. The sEVs CD63+ percentage from ZIKV-infected cultures (19.32 ± 0.93%) was 1.7-fold higher than the sEVs CD63+ from the mock cultures (11.08 ± 0.34%), showing a significant difference (*p* < 0.01) (Figure 3D,E). The MFI values (obtained by FACS) were proportional to the sEVs CD63+ presence (Figure S9D). We found that the sEVs CD63+ from ZIKV-infected cells showed MFI values 1.3-fold higher than the sEVs CD63+ from the mock cultures.

The TEM images (Figure 3F) from the mock and ZIKV-infected C6/36 sEVs (500 nm scale) show a heterogeneous population of sEVs [46] in terms of their size (fewer than 200 nm in diameter), shape, and content; this population is also well defined by a bilipid membrane. We did not identify viral particles inside sEVs TEMs from ZIKV-infected cells, so we evaluated the ZIKV E protein presence on the membrane's surface of sEVs. We found 31.19 ± 0.28% of positive ZIKV E protein+ sEVs coupled with paramagnetic nanobeads (*p* < 0.001) compared with the sEVs from the mock C6/36 cells (Figure 3G,H). The ZIKV E protein (MFI) in the sEVs from ZIKV-infected cells was 4.03-fold higher than that of the sEVs from the mock cultures (Figure S9E).

**Figure 3.** ZIKV-infected C6/36 cells issue small EVs (sEVs) CD63+. (**A**) CD63-like detection at 24, 48, 72, 96, and 120 h PI by FACS. Dot plots are the representative mean ± SD of positive cells from three independent experiments. Isotype IgG1 antibody was used as negative control. (**B**) Cells CD63-like+ percentages obtained by FACS. The levels of CD63-like protein were compared (by an unpaired Student's t-test) with the isotype control (+) and the mock cells (\*) values. (**C**) NTA of the purified sEVs isolates from the mock and the ZIKV-infected C6/36 cells. Histograms are the representative mean ± SD of the nanoparticle's concentration (particles/mL) and the size (nm) from three independent experiments. (**D**) sEVs CD63+ coupled with paramagnetic bead detection by FACS. Dot plots are the representative mean ± SD of the sEVs CD63+ from three independent experiments. (**E**) sEVs CD63+ percentages obtained by FACS. The CD63 levels were compared (by an unpaired Student's t-test) with the mock sEVs CD63+ (\*) values. (**F**) Transmission electron microscopy (TEM) images from the mock and the ZIKV-infected C6/36 sEVs (500 nm scale). (**G**) ZIKV E protein detection on the sEVs coupled with paramagnetic beads by FACS assay. Dot plots are the representative mean ± SD of the positive sEVs ZIKV E protein+ from three independent experiments. (**H**) Percentages of the sEVs ZIKV E protein+ by FACS. The ZIKV E protein levels were compared (by an unpaired Student's t-test) with the sEVs mock C6/36 (\*) values. Statistical significance was recognized as ++ or \*\* when *p* < 0.01, and +++ or \*\*\* when *p* < 0.0001.

#### *3.4. ZIKV C6*/*36 EVs, after ZIKV Inactivation, Carry Viral RNA, Reproduce Lytic Plaque Formation on Vero Cells, and Favor Infection in Naïve Mosquito Cells*

We determined whether ZIKV mosquito EVs participate during the infectious process as was recently reported for other arthropod-borne flaviviruses [16,30,36]. It was then determined whether ZIKV C6/36 EVs contain viral elements as RNA or E protein and whether they support the infection of naïve cells.

First, we determined whether ZIKV could be inactivated by RNAse A activity and UV radiation at 1200 μJ (× 100) in three consecutive cycles (Figure S4A). We evaluated different conditions for incubation times and found that the complete RNA degradation from ZIKV viral stock occurred for 1 h at 37 ◦C (Figure 4A). The infection capability of inactivated ZIKV (iZIKV) was evaluated by a lytic plaque assay (Figure 4B) and found no lytic plaque formation after the inactivation process.

We then proceeded to inactivate free ZIKV virions in C6/36 mosquito EV isolates (small/large). These samples were irradiated at 1200 μJ (×100) in three consecutive cycles by using an UV Stratalinker (Figure S4B). In addition, the possible presence of ZIKV genomic RNA in the EVs samples, as a possible precipitation product during the EVs isolation process, was eliminated with RNase A, as described above: By using 10 μg/mL of RNase A (DNase and Protease-free) added to the different EV isolates for 1 h at 37 ◦C, we observed total RNA degradation in all samples (sEVs and lEVs) (Figure 4A).

To evaluate the integrity of the EVs (sEVs and lEVs) after the ZIKV inactivation process, we proceed to quantified them by NTA. The NTA from lEVs ZIKV C6/36 (RNase A + UV) showed a concentration of 2.48 × 10<sup>10</sup> ± 4.07 × 10<sup>9</sup> particles/mL with an average size value of 304.1 ± 10.9 nm, while the NTA from sEVs ZIKV C6/36 (RNase A + UV) showed a concentration of 2.49 × 10<sup>11</sup> ± 2.29 × 10<sup>10</sup> particles/mL with an average size value of 150.9 ± 5.5 nm (Figure 4C). The NTA histograms showed the same patterns of the NTA histograms shown in Figures 2A and 3C, so the EVs' integrity is preserved. In Figure 4C, the peak containing particles of nearly 50 nm, compatible with ZIKV, was substantially reduced.

Next, to evaluate the possible ZIKV genomic RNA presence inside small and large EVs (RNase A + UV-treated), all samples were processed for RNA extraction and purification by using the QIAamp RNA Mini kit, according to the manufacturer's instructions. The samples were first lysed under highly denaturing conditions using the lysis buffer. The ZIKV-RNA amplification was performed by RT-PCR, according to the specifications given by the OneStep RT-PCR kit (see Materials and Methods). The amplified cDNA corresponded to the specific 364 bp E-amplicon for the ZIKV [40] envelope protein, which was visualized on 2% ethidium bromide-stained 1.2% agarose gel (Figure 4D). These findings sugges<sup>t</sup> that both small and large EVs from ZIKV-infected mosquito cells may carry viral RNA after RNase A + UV treatment.

As a result of the significance of these data, we also evaluated the possible mammalian naïve cell infection via small/large ZIKV C6/36 EVs. With this aim, first, we used the epithelial cells (Vero) from the monkey *Cercopithecus aethiops* (used as gold standard cells for infection assay) [32] to perform plaque assays (as described above) in the presence of small and large ZIKV C6/36 EVs (also RNase A + UV-treated and untreated samples). The presence of lytic plaques in ZIKV-infected Vero cells was present in high or undetermined amounts at different dilutions (Figure 4E). Importantly, lytic plaques were also observed in cultures in the presence of small and large ZIKV C6/36 EVs isolates in a more concentrated amount, which were also formed in high or undetermined quantities. However, plaque formation was not detected in the negative control named non-EVs ZIKV SNT (the final supernatant obtained in the last centrifugation during the isolation of EVs from ZIKV-infected C6/36 cell culture media). These data sugges<sup>t</sup> the ZIKV infection of monkey epithelial cells via lEVs/sEVs released from ZIKV-infected C636 mosquito cells.

Therefore, different EVs stimulation assays were performed using naïve C6/36 cells in the presence of ZIKV (MOI 1), mock C6/36 EVs, or ZIKV C6/36 small/large EVs (including RNase A + UV-treated and untreated isolates) (Figure S5). The following conditions were applied: 0.10 mg of protein in 250 μL/well of EVs isolates from mock C6/36 cells (lEVs and sEVs, separately), 0.10 mg of protein in 250 μL/well of EVs isolates from ZIKV-infected C6/36 cells (lEVs and sEVs, separately), 0.10 mg of protein in 250 μL/well of EVs isolates from ZIKV-infected C6/36 cells RNase A + UV-treated (lEVs

and sEVs, separately), and 0.10 mg of protein in 250 μL/well of non-EVs ZIKV SNT. The mock cells and ZIKV-infected cells (MOI 1) were used as negative and positive controls, respectively. We also evaluated the iZIKV infection capability on C6/36 cells (Figure S11A,B), and we found that iZIKV did not infect naïve C6/36 cells after 48 h of incubation, because the ZIKV E protein was undetectable.

**Figure 4.** EVs from ZIKV-infected C6/36 cells, after ZIKV inactivation by RNase A activity assay and UV radiation, carry viral RNA and favor the mammalian cell infection. (**A**) RNase A activity assay. RNase A (10 μg/mL) was added to the purified ZIKV RNA and incubated for 1 h, 30 min, and 15 min at 37 ◦C with 5% CO2. Additionally, the RNase A activity (1 h incubation) was evaluated in C6/36 EVs isolates. The RNA degradation pattern was visualized on 2% ethidium bromide-stained 1.2% agarose gel. (**B**) Inactivated ZIKV (iZIKV) titration by a lytic plaque assay. (**C**) NTA of the EV isolates (from ZIKV-infected cells) treated with RNase A and UV. Histograms are the representative mean ± SD of the nanoparticle's concentration (particles/mL) and the size (nm) from three independent experiments. (**D**) ZIKV RNA detection (RT-PCR) in ZIKV-infected C6/36 EVs (RNase A + UV-treated) samples. The ZIKV amplicon (364 bp from E genome conserved region) was visualized on 2% ethidium bromide-stained 1.2% agarose gel. (**E**) Evaluation of the ZIKV-infected C6/36 EVs (RNase A + UV-treated and untreated samples) in the viral transmission to naïve Vero cells by a lytic plaque assay.

The samples were incubated over 48 h (the best infection time) and evaluated for possible viral infection by means of ZIKV E protein detection using a FACS assay. The ZIKV E protein presence was detected in 44.39 ± 0.69% of ZIKV-infected C6/36 cells (Figure 5A–C), showing statistical significance (*p* < 0.0001) in relation to the mock cells (\*) (34.7-fold higher). In naïve C6/36 cells stimulated with ZIKV C6/36 lEVs, the E protein was found in 38.24 ± 1.15% of cells and 28.44 ± 1.37% of cells stimulated with ZIKV C6/36 lEVs (RNase A + UV-untreated and treated) (#,!), showing statistical significance (*p* < 0.0001) compared with the mock C6/36 cells and stimuli in the presence of mock C6/36 EVs (Figure 5A,B). Similarly, in naïve mosquito cells cultured with ZIKV C6/36 sEVs, the viral E protein was found in 39.31 ± 0.58% of cells and 33.00 ± 0.29% of cells stimulated with ZIKV C6/36 sEVs (RNase A + UV-untreated and treated) (/, ~), showing statistical significance (*p* < 0.0001) compared with the mock cells and naïve cells stimulated with EVs from mock C6/36 cells (Figure 5A–C). The present data sugges<sup>t</sup> that ZIKV C6/36 (large and small) EVs could favor the infection of mosquito naïve cells.

**Figure 5.** ZIKV E protein is present on the membrane's surface of naïve C6/36 cells after the stimulus with ZIKV-infected C6/36 EVs. (**A**) ZIKV E protein detection at different EV stimuli conditions (FACS assay). Dot plots are the representative mean ± SD of the positive cells from three independent experiments. (**B**) Percentages of ZIKV E protein+ cells (FACS) after the lEVs stimuli. The levels of the ZIKV E protein were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, +, #, !, or ◦ when *p* < 0.05, \*\*, ++, ##, !!, or ◦◦ when *p* < 0.01, and \*\*\*, +++, ###, !!!, or ◦◦◦ when *p* < 0.0001. (**C**) Percentages of ZIKV E protein+ cells (FACS) after the sEVs stimuli. The levels of the ZIKV E protein were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, ¡,/, or ~ when *p* < 0.05, \*\*, ¡¡,//, or ~~ when *p* < 0.01, and \*\*\*, ¡¡¡,///, or ~~~ when *p* < 0.0001.

Likewise, in parallel assays, all samples were evaluated for the presence of ZIKV E protein by fluorescence microscopy and by the cytopathic effects observation using light field microscopy (Figure

S12). The cytopathic e ffects in the light fields were indicated with black arrows (20×). An increased cytopathic e ffect with the formation of syncytium structures was present in the ZIKV-infected cultures but also in cultures stimulated with small and large ZIKV C6/36 EVs (RNase A + UV-treated and untreated) isolates. The ZIKV E protein (red) was detected in naïve C6/36 cells stimulated with the small and large ZIKV C6/36 EVs at similar levels and patterns of ZIKV-infected mosquito cells (fluorescence microscopy, 60×) and were undetected in cultures of naïve C6/36 cells in the presence of mock C6/36 EVs (small and large). The same was found for the non-EVs ZIKV supernatant or mock C6/36 cultures.

#### *3.5. ZIKV-Infected C6*/*36 EVs Participate during Infection of Naïve Human Monocytes*

Recently, it was reported that dengue virus (DENV) uses small EVs of C6/36 mosquito cells for its transmission from the vector to mammalian host cells, including human skin keratinocytes and ECs [16,30]. However, to date, it has not been determined if EVs from ZIKV-infected C6/36 cells participate during monocyte infection. As monocytes are the main target cells during ZIKV human host infection [17,47], we evaluated the possible participation of large and small EVs in the potential. Initially, we evaluated the viral infection of human monocytes (MOI 1) via the detection of the viral E protein at the surfaces of the membrane monocytes at 24, 48, 72, 96, and 120 h PI using a FACS assay (Figure 6).

**Figure 6.** ZIKV (MOI 1) infects human monocytes. ( **A**) ZIKV E protein detection at 24, 48, 72, 96, and 120 h PI by FACS assay. Dot plots are the representative mean ± SD of the positive cells from five independent experiments. (**B**) ZIKV-infected cells percentages obtained by FACS assay. The ZIKV E protein levels were compared (by an unpaired Student's t-test) with mock C6/36 (\*) values. Statistical significance was recognized as \*\*\* when *p* < 0.0001.

ZIKV was able to establish a productive infection of human monocytes, since the viral E protein was detected at high levels on the cell membrane's surface in all post-infection time points of the assay (Figure 6A). However, the greater percentages of ZIKV-positive cells were present at 24 h PI (79.98 ± 1.07%), at 96 h PI (77.92 ± 1.17%), and at 120 h (78.01 ± 0.98%), which were statistically significant (*p* < 0.0001) when compared to the mock cells. Likewise, we found that the ZIKV infection of THP-1 human monocytes favor progressive activation and cell di fferentiation e ffects (Figure S13A). The cytopathic effect observation by light field microscopy supports the presence of higher amounts of adherent cells between 96 and 120 h PI. Consequently, we decided to evaluate the ZIKV infection of naive monocytes after the EVs stimulation at 96 h PI.

Di fferent EVs stimulation assays were performed using naïve THP-1 cells in the presence of ZIKV (MOI 1), mock C6/36 EVs, or ZIKV C6/36 EVs (small and large isolates as the same for the RNase A + UV-treated and untreated samples) (Figure S5), which were evaluated for possible infection by means of ZIKV E protein detection by FACS (see Materials and Methods). We also evaluated the

iZIKV infection capability on THP-1 cells (Figure S11C,D) and found that iZIKV did not infect naïve monocytes after 96 h of incubation, because the ZIKV E protein was undetectable.

As was expected, the viral E protein was detected in 73.41 ± 0.59% of ZIKV-infected monocytes and was statistically significant (*p* < 0.0001) when compared to the mock cells (\*) (Figure 7A–C). The ZIKV E protein was also detected in higher amounts on naïve THP-1 cells that were stimulated with ZIKV C6/36 lEVs (54.89 ± 0.56% in RNase A + UV-treated (!) isolates and 70.23 ± 1.53% in untreated (#) isolates) and ZIKV C6/36 sEVs (38.29 ± 0.79% in RNase A + UV-treated (~) isolates and 41.34 ± 0.39% in untreated (/) isolates). The ZIKV E protein levels were statistically significant (*p* < 0.0001) when compared to the ZIKV-C6/36 EVs stimulated cultures against the mock cells, mock C6/36 EVs, and non-EV ZIKV SNT (◦) cultures (Figure 7A–C).

**Figure 7.** ZIKV E protein is present on the membrane's surface of naïve monocytes (THP-1 cells) after the stimulus with ZIKV-infected C6/36 EVs. (**A**) ZIKV E protein detection at different EV stimuli conditions (FACS assay). Dot plots are the representative mean ± SD of the positive cells from three independent experiments. (**B**) Percentages of ZIKV E protein+ cells (FACS) after the lEVs stimuli. The levels of the ZIKV E protein were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, +, #, !, or ◦ when *p* < 0.05, \*\*, ++, ##, !!, or ◦◦ when *p* < 0.01, and \*\*\*, +++, ###, !!!, or ◦◦◦ when *p* < 0.0001. (**C**) Percentages of ZIKV E protein+ cells (FACS) after the sEVs stimuli. The levels of the ZIKV E protein were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, ¡,/, or ~ when *p* < 0.05, \*\*, ¡¡,//, or ~~ when *p* < 0.01, and \*\*\*, ¡¡¡,///, or ~~~ when *p* < 0.0001.

The present data sugges<sup>t</sup> that ZIKV-infected C6/36 EVs not only support the infection of naïve mosquito cells but also participate during infection of mammalian host cells, as in the case of human monocytes, which are important immune effector cells during host–pathogen interplay.

#### *3.6. ZIKV-Infected C6*/*36 EVs Promote Change in Monocyte Phenotype (CD14, CD16, and CD11b)*

It is known that EVs may have di fferent functions when interacting with other cells, modifying their naïve cellular behavior [48]. Therefore, if EVs released by ZIKV-infected mosquito cells were able to infect human monocytes, they could also favor monocyte activation and/or di fferentiation. To assess this objective, naïve human monocytes were stimulated by the presence of ZIKV (MOI 1), mock C6/36 EVs, or ZIKV C6/36 EVs (small and large isolates as the same for the RNase A + UV), which were evaluated for monocyte activation or di fferentiation (to adherent phenotype cells) by means of the monocytes' phenotypic shift from a classical (CD14++ CD16−) to an intermediate (CD14++ CD16+) or non-classical (CD14+ CD16++) phenotype, which seems to be the main producer of inflammatory mediators in response to viral infection [49].

By the stimulation assays in the presence of the di fferent EVs samples, we observed (Figure 8A,B) that the classical monocyte phenotype changes to CD14++ CD16+ intermediate monocytes with respect to the mock THP-1 cells and those stimulated with the mock C6/36 EVs (*p* < 0.0001 for CD14 and *p* < 0.01 for CD16). Moreover, in a parallel assay, we observed that naïve monocytes were di fferentiated and expressed CD11b+ at the membrane's surface (Figure 8C) with elevated levels in ZIKV-infected monocytes and those stimulated with ZIKV C6/36 EVs (*p* < 0.0001) compared with the mock THP-1 cells and those stimulated with mock C6/36 EVs. On the other hand, naïve monocyte activation by the ZIKV C6/36 EVs was also observed by light-field microscopy (Figure S13B), showing the transformation of naïve cells to the adherent phenotype (black arrows), as similar levels occur in ZIKV-infected monocytes.

It has been suggested that monocyte subsets (intermediate and non-classical) play an essential role in immunopathology during *Flavivirus* infection [49], since non-classical monocytes seem to be the main producers of pro-inflammatory mediators in response to viral infection. The present data show that ZIKV C6/36 (small/large) EVs activate and di fferentiate naïve monocytes, changing cells to a pro-inflammatory state, in a similar mode to ZIKV (MOI 1) infection.

Next, to evaluate the possible expression of a pro-inflammatory response induced by EVs released from ZIKV-infected C6/36 cells, we performed a detection of the tumor necrosis factor-alpha (TNFα) mRNA expression in naïve monocytes infected by ZIKV (MOI 1), and those stimulated with the mock C6/36 EVs or ZIKV C6/36 EVs (RNase A + UV-treated and untreated isolates). We found that, like ZIKV-infected monocytes, ZIKV C6/36 sEVs were able to induce TNFα mRNA expression in human monocytes (Figure 8D). TNFα is a pro-inflammatory cytokine that was recently determined to be an important host factor involved in neurological disorders and central nervous system inflammation during ZIKV human infection [50].

A growing number of evidence indicates that sEVs are involved in inflammatory processes or immune responses that play an important role in a large number of pathologic states, including infectious diseases. sEVs can modulate gene expression and the functions of the cells with which they interact, and their content depends on the cells from which they are released [23,27]. We found that sEVs from ZIKV-infected C6/36 cells induce immunophenotype changing (to intermediate/non-classical) in monocytes. This proinflammatory phenotype could be directly implicated in TNFα mRNA expression.

**Figure 8.** EVs from ZIKV-infected C6/36 favor the pro-inflammatory phenotype change in naïve human monocytes. (**A**) Monocytes CD14+ percentages (FACS) at different EV stimuli conditions from three independent experiments. (**B**) Monocytes CD16+ percentages (FACS) at different EVs stimuli conditions from three independent experiments. (**C**) Monocytes CD11b+ percentages (FACS) at different EVs stimuli conditions from three independent experiments. The CD14, CD16, or CD11b levels were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, +, ¡, #, !,/, ~, or ◦ when *p* < 0.05, \*\*, ++, ¡¡, ##, !!,//, ~~, or ◦◦ when *p* < 0.01, and \*\*\*, +++, ¡¡¡, ###, !!!,///, ~~~, or ◦◦◦ when *p* < 0.0001. (**D**) Tumor necrosis factor-alpha (TNF-α) mRNA expression (RT-PCR) in naïve monocytes at different EVs stimuli conditions. The TNF-α genome conserved region (amplicon of 600 bp) was visualized on 2% ethidium bromide-stained 1.2% agarose gel.

#### *3.7. ZIKV C6*/*36 EVs Participate during Infection of Naïve Endothelial Vascular Cells*

The ZIKV mosquito EVs participation during infection of vascular ECs is unknown. We first determined whether ZIKV (MOI 1) was able to infect human endothelial vascular cells, by means of the viral E protein detection on cells membrane surface at 24, 48, 72, 96, and 120 h PI by the FACS assay. First, the optimal conditions for ZIKV vascular endothelial cell infection were evaluated (Figure 9).

**Figure 9.** ZIKV (MOI 1) infects human endothelial (HMEC-1) cells. (**A**) ZIKV envelope (E) protein detection at 24, 48, 72, 96, and 120 h PI by the FACS assay. Dot plots are the representative mean ± SD of the positive cells from five independent experiments. (**B**) ZIKV-infected cells percentages obtained by FACS. The ZIKV E protein levels were compared (by an unpaired Student's t-test) with the mock C6/36 (\*) value. Statistical significance was recognized as \* when *p* < 0.05, \*\* when *p* < 0.01, and \*\*\* when *p* < 0.0001.

ZIKV was able to establish a productive infection in microvascular endothelial (HMEC-1) cells (Figure 9A,B), since viral E protein was detected at high levels on the cell membrane surface mainly at 24 (45.62 ± 1.76%) and 72 h (38.22 ± 1.20%) PI time points, with a significance level of *p* < 0.0001 when compared to the mock HMEC-1 cells (22.9- and 19.2-fold higher, respectively). Likewise, when these samples were observed by light-field microscopy, the ZIKV-infected ECs showed important cytopathic effects (black arrows), with a formation of syncytium structures at 48 and 72 h PI, but without monolayer detachment (Figure S14A). We also evaluated the iZIKV infection capability on HMEC-1 cells (Figure S11E,F) and found that iZIKV did not infect naïve ECs after 72 h of incubation, because the ZIKV E protein was undetectable.

Afterward, the possible participation of small/large C6/36 EVs in the potential infection of vascular ECs was evaluated. Next, different stimulation assays were performed using 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 same for the RNase A + UV-treated and untreated). All stimuli were evaluated by measuring the ZIKV E protein presence by FACS (Figure 10).

As shown in Figure 10A, viral E protein was present at high levels in ZIKV (MOI 1) infected HMEC-1 cells (30.22 ± 0.58%), but also at high levels in naïve ECs stimulated by the ZIKV C6/36 sEVs (RNase A + UV-treated (19.28 ± 0.80%) and untreated (26.40 ± 0.78%)) and ZIKV C6/36 lEVs (RNase A + UV-treated (15.28 ± 0.49%) and untreated (26.32 ± 0.56%)). The percentage of viral E protein was compared between all conditions' values against the mock HMEC-1; these values were statistically significant (*p* < 0.0001) (Figure 10B,C). These data sugges<sup>t</sup> that ZIKV infected-C6/36 EVs (large and small) support the infection of mammalian host cells, including vascular ECs.

**Figure 10.** ZIKV E protein is present on the membrane's surface of naïve endothelial cells (HMEC-1) after the stimulus with ZIKV-infected C6/36 EVs. (**A**) ZIKV E protein detection at different EVs stimuli conditions (FACS assay). Dot plots are the representative mean ± SD of the positive cells from three independent experiments. (**B**) Percentages of ZIKV E protein+ cells (FACS) after the lEVs stimuli. The levels of the ZIKV E protein were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, +, #, !, or ◦ when *p* < 0.05, \*\*, ++, ##, !!, or ◦◦ when *p* < 0.01, and \*\*\*, +++, ###, !!!, or ◦◦◦ when *p* < 0.0001. (**C**) Percentages of ZIKV E protein+ cells (FACS) after the sEVs stimuli. The levels of the ZIKV E protein were compared (by an unpaired Student's t-test) between all conditions' values. Statistical significance was recognized as \*, ¡,/, or ~ when *p* < 0.05, \*\*, ¡¡,//, or ~~ when *p* < 0.01, and \*\*\*, ¡¡¡,///, or ~~~ when *p* < 0.0001.

#### *3.8. ZIKV C6*/*36 EVs Favor a Pro-Inflammatory and Pro-Coagulant State of Vascular Endothelial Cells and Promote the Endothelial Vascular Cells' Permeability*

Recent observations in humans and animal models [51,52] sugges<sup>t</sup> that, in severe Zika cases, different coagulation disorders occur. It has been shown that some viruses activate the coagulation system through tissue factor (TF) receptor expression [53]. We previously reported that the DENV 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, induce signaling inflammatory pathways (via phosphorylation of MAPKs p38 and ERK1/2, by transcription of NF-κB factor), thereby supporting the upregulation of VCAM-1 adhesion or pro-inflammatory molecules in ECs [54]. Next, we determined whether the ZIKV infection (MOI 1) of vascular endothelial cells (HMEC-1) or the stimulation by ZIKV C6/36 EVs favor a pro-inflammatory/pro-coagulant state of naïve endothelial vascular cells.

Therefore, we assessed the possible participation of small and large EVs issued from ZIKV-infected C6/36 cells for the induction of coagulation (TF) or inflammation (PAR-1) receptors at the membrane's surface of ECs. In a parallel assay, the adhesion ICAM-1 molecule was also evaluated. Different stimulation assays were performed using naïve HMEC-1 cells in the presence of ZIKV (MOI 1), the mock C6/36 EVs, or the ZIKV C6/36 EVs (small and large isolates, the same for the RNase A + UV), which were evaluated for the presence of TF, PAR-1, and ICAM-1 at their cell membranes (FACS) at 72 h post-stimulus (Figure 11A–C). Likewise, the cytopathic effect formation on these samples was evaluated by light-field microscopy (Figure S14B).

Elevated levels of TF (20.27 ± 0.51%) were detected in ZIKV-infected (MOI 1) vascular ECs (Figure 11A), but also in the presence of large ZIKV C6/36 EVs (16.75 ± 0.59%), which were both statistically significant (*p* < 0.0001) when compared with the mock HMEC-1 cultures (\*). In small EV culture samples, the TF values were 4.82 ± 0.23%. The upregulation of the TF receptor may trigger the generation of hemostatic proteases (thrombin) favoring the activation of protease-activated receptors (PARs). Figure 11B shows the activation percentages of the PAR-1 in ECs infected with ZIKV (26.41 ± 0.56%), and the same is true for large ZIKV C6/36 EVs (23.17 ± 0.33% for untreated and 13.15 ± 0.38% for RNase A + UV-treated) and small ZIKV C6/36 EVs (21.55 ± 0.14% for untreated and 16.19 ± 0.79% for RNase A + UV-treated) culture samples. We found a significant difference (*p* < 0.0001) in all EVs stimuli compared with the mock HMEC-1 cultures. It is well known that PAR-1 favors signaling pathways for the expression of pro-adherent and pro-inflammatory molecules [55]. Therefore, we assessed ICAM-1 (Figure 11C) detection on EC surfaces in ZIKV-infected HMEC-1 cultures (12.87 ± 0.16%). The same was observed in cultures of naïve ECs in the presence of the mock C6/36 EVs or the ZIKV C6/36 EVs (small/large isolates, the same for the RNase A + UV) (*p* < 0.001). These data were corroborated by light-field microscopy (Figure S14B), where ZIKV-infected ECs showed an increased cytopathic effect (black arrows) with the formation of vacuolization and syncytial structures. Cytopathic effects were also observed in all naïve EC culture assays stimulated by the ZIKV C6/36 EVs.

Moreover, to evaluate the possible expression of the pro-inflammatory response by the stimulation of EVs released from ZIKV-infected C6/36 cells, we measured the TNF-α mRNA expression in naïve ECs infected by ZIKV (MOI 1) and those stimulated by the mock C6/36 EVs and the ZIKV C6/36 EVs (small/large isolates as the same for the RNase A + UV). We found that, similar to ZIKV infection, ZIKV C6/36 EVs were able to induce TNF-α mRNA expression in endothelial vascular cells (Figure 11D). Our data sugges<sup>t</sup> the possible participation of the coagulation–inflammation process in the coagulation disorders present in severe cases of Zika. The endothelial vascular cell activation (damage) during ZIKV infection with an inflammatory response can cause EC dysfunction and weaken the endothelial barrier integrity. Thus, we evaluate the vascular endothelial barrier integrity in vitro using a Transwell assay (Figure 11E).

Our data indicate that endothelial vascular cells are susceptible to ZIKV infection and activation by both ZIKV (MOI 1) and ZIKV (small/large) C6/36 EVs with pro-inflammatory cytokine expression, which could increase endothelial monolayer permeability. Therefore, we determined whether ZIKV infection or ZIKV C6/36 EVs disturb the vascular endothelial barrier's integrity in vitro using a Transwell assay (see Materials and Methods; Figure S6), performed in a naïve EC culture in the presence of ZIKV (MOI 1), the mock C6/36 EVs, or the ZIKV C6/36 EVs.

For ZIKV-infected ECs, the permeability percentage (15.77 ± 1.23%) increased 2.3-fold compared to the mock HMEC-1 cultures (*p* < 0.0001). In the presence of the ZIKV C6/36 lEVs (RNase A + UV-treated (10.87 ± 1.29%) and untreated (10.95 ± 1.37%)), permeability increased, on average, 1.6-fold (*p* < 0.05); in cultures stimulated by ZIKV C6/36 sEVs (RNase A + UV-treated (10.29 ± 1.14%); when untreated (12.33 ± 1.19%)), the endothelial permeability increased 1.5- (*p* < 0.05) and 1.8-fold (*p* < 0.01), respectively (Figure 11E). These data sugges<sup>t</sup> that ZIKV C6/36 EVs may participate in vascular endothelial damage with a weakening of the endothelial barrier integrity and support the mosquito EVs participation during the infection process, which could contribute to the pathogenesis of ZIKV infection in a human host.

**Figure 11.** EVs from ZIKV-infected C6/36 modify towards a pro-coagulant, pro-inflammatory, and pro-adherent phenotype and favor permeability in naïve endothelial cells (ECs). (**A**) EC TF-1+ percentages (FACS) at different EVs stimuli conditions from three independent experiments. (**B**) EC Protease Activated Receptor+ (PAR-1) percentages at different EVs stimuli conditions from three independent experiments. (**C**) EC intercellular adhesion molecule-1+ (ICAM-1) percentages at different EVs stimuli conditions from three independent experiments. The TF, PAR-1, or ICAM-1 levels were compared (by an unpaired Student's t-test) between all conditions' values. (**D**) TNF-α mRNA expression (RT-PCR) in naïve ECs at different EVs stimuli conditions. The TNF-α genome conserved region (amplicon of 600 bp) was visualized on 2% ethidium bromide-stained 1.2% agarose gel. (**E**) Permeability percentages obtained by assessing the fluorescein isothiocyanate (FITC)-Dextran pass through the EC monolayers in the presence of different EV stimuli conditions. Three independent experiments were performed. For the 100% FITC-Dextran delivered control, a no-cell insert was used. The endothelial vascular permeability percentages were compared (by an unpaired Student's t-test) between all conditions' values. For all experiments, statistical significance was recognized as \*, +, ¡, #, !,/, ~, or ◦ when *p* < 0.05, \*\*, ++, ¡¡, ##, !!,//, ~~, or ◦◦ when *p* < 0.01, and \*\*\*, +++, ¡¡¡, ###, !!!,///, ~~~, or ◦◦◦ when *p* < 0.0001.
