Persistent and Long-Term Infectivity of Dengue Virus in Mosquito Cells Revealed Reduced Replication in Vector Host and Human Endothelial Cells

Abstract

Understanding the intrinsic potential of persistent dengue virus (DENV) replication and survival in vector host cells is critically important. In this study, we investigated to what extent DENV can replicate within the vector host Aedes albopictus C6/36 mosquito cells (cell line routinely used for propagation of DENV in research laboratories). We detected DENV serotype 2 (DENV2) loads in cell culture supernatants collected at different days post infection (3, 19, 33, 60, 90, 120 and 175) and found the presence of capsid transcripts and protein levels in these virus supernatants. Tissue culture infectious dose 50 (TCID50) assay revealed a gradual reduction in viral titers and infectivity from days 19 to 175 post DENV2 infection. Furthermore, infection kinetics with these virus supernatants collected at different days post DENV2 infection demonstrated declining viral replication in naïve C6/36 cells and human endothelial recipient cells. These results provided information on viral replication competence and the persistency of DENV2 infection from days 19 to 175 in mosquito cells. Extracellular vesicles (EVs) isolated from DENV2-infected C6/36 cell culture supernatants showed a progressive increase in EV concentration from day 33 to day 175. While DENV2 loads within these EVs declined over time, their ability to mediate infection in naïve C6/36 and endothelial cells remained constant. Notably, the viral membrane (M) protein was detected in EVs at days 3, 19, and 33 but was absent at later timepoints (days 60, 90, 120, and 175). The prM protein was not detected in any of the samples analyzed. In conclusion, DENV2 exhibits the capacity for persistent infection in mosquito cells, thereby potentially serving as a model for investigating the mechanisms that govern years of long-term and sustained viral infections within the vector host.

1. Introduction

DENV comprises four serotypes (1–4), belongs to the family of Flaviviridae and is transmitted primarily by Aedes sp. mosquitoes [1,2,3,4,5]. According to the World Health Organization (WHO), mosquito-borne dengue transmission imposes a significant public health concern, and the outcome challenges suggest this flavivirus to be an emerging pathogen. Homologous infections with any of the DENV serotypes typically resolves within 7 to 10 days post infection (p.i.) due to natural host immunity, conferring long-term protection against the specific serotype [4,6,7,8,9]. Approximately 390 million cases of DENV infection cases are reported annually, with 20–30% of patients exhibiting high fever and flu-like symptoms [2,4,8,10,11,12]. Notably, DENV2-mediated heterologous infection is the most prevalent and associated with severe outcomes, including dengue fever and dengue hemorrhagic fever, resulting in an estimated 25,000 deaths annually [9,13]. Despite the significant burden of DENV, the effective drugs or vaccines available against its infection require further improvement [7,8,14,15,16,17].

DENV is a positive-sense single-stranded RNA virus with a genome consisting of 11 kilobase pairs, encoding three structural proteins (C, prM, and E) and seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5), and all proteins overhang a single open reading frame (ORF) [4,18,19]. We previously reported the presence of a full-length DENV2 viral RNA genome inside the extracellular vesicles (EVs) derived from infected C6/36 mosquito cells [17]. The presence of an entire DENV2 RNA genome in mosquito EVs is sufficient for infection and EVs from infected mosquito cells are highly infectious to the vertebrate host [17]. Evaluating the infection biology, DENV enters the female Aedes aegypti or Aedes albopictus mosquito body through a blood meal from an DENV-infected vertebrate host including humans [2,5,7,12,15,18,20,21,22]. DENV replicates inside the mid-gut of the mosquitoes, crosses the epithelium barrier and localizes or resides inside the salivary glands. Surprisingly, newly released virus particles circulate throughout the vector body and then enter into the salivary glands for transmission [7,22]. Though A. aegypti and A. albopictus mosquitoes exhibit several antimicrobial peptides (AMPs), RNA interference (RNAi) and antiviral pathways combat the virus propagation within their bodies, but these flaviviruses still securely reside in the vector host for a long period of time and until the next blood meal feeding [5,8,18,21,22,23]. In addition, clinical case reports from DENV-infected human patients have documented the following points: (i) acute or subacute neurological complications temporarily associated with dengue infection or (ii) persistent symptoms with long-term neurological sequelae that would last from a few months to a couple of years after symptomatic illness or in exceptional cases of immunocompromised conditions that may have prolonged viral detection [24,25,26,27,28,29,30,31,32,33,34,35,36,37,38]. Certainly, the concept of DENV entering a latent state within the human host for prolonged durations remained negligible. Moreover, the existence of an appropriate animal model for investigating the ability of long-term DENV infectivity and its maintenance and propagation is essential but there are no studies reporting such advancement(s) in the field. While long-term and persistent DENV infection in human mononuclear cell lines has been well-documented [39] with the appearance of DENV pathogenesis, investigations into persistent infection within the mosquito host remain unexplored. A study involving serial passaging of DENV in mosquito cells for up to 56 weeks (or about 1 year) revealed that virions released from persistently infected cells retained infectivity in mosquitoes but not in vertebrate host cells [4,5,18,20]. Our current findings underscore the importance of elucidating this long-term and persistent DENV2 infection in mosquito cells, which may encourage the development of effective therapeutics in the future [2,8,14,15,16]. Therefore, it is critically important to elucidate the infectivity of DENV2 (viral supernatants collected from cells at 3, 19, 33, 60, 90, 120, and 175 days post-infection) and understand its replication process in vector host mosquito cells and in human endothelial cells.

2. Materials and Methods

2.1. Cells, Virus Supernatant Preparations, and Infection of Naïve Recipient Cells

C6/36 cells (Aedes albopictus cell line, catalog number CRL 1660) were obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA) and cultured in complete EMEM media and according to the ATCC’s guideline(s) at 28 °C and 5% CO₂. EMEM media was supplemented with 10% heat-inactivated fetal bovine serum (FBS obtained from VWR, Radnor, PA, USA), 1% penicillin/streptomycin, glutamine mixture and amphotericin B mixture (obtained from Thermo/Fisher Scientific, Waltham, MA, USA). EA.hy926 cells (human endothelial cells, catalog number CRL-2922) were obtained and maintained in DMEM media (from ATCC, Manassas, VA, USA), supplemented with 10% heat-inactivated FBS, 1% penicillin/streptomycin, and glutamine mixture by following the company’s guidelines (cultured at 37 °C in the presence of 5% CO₂). DENV (serotype 2) New Guinea C (NGC) strain, catalog number NR-84, obtained from BEI resources (Manassas, VA, USA) was propagated in C6/36 cells cultured in a T75 flask and as previously described [40]. DENV2-infected mosquito cells were maintained continuously in the same culture flask, but cells were replaced with fresh complete media after each timepoint of virus supernatant collection. Mosquito C6/36 cells were infected with unknown titers of DENV2 (obtained from laboratory stock), and the cell culture supernatants were collected after prolonged DENV2 infection and without any bias. No new growing cells were added to the persistently infected mosquito cells. Each of these mosquito cell culture supernatants collected at different days post DENV2 infection were randomly selected based on the need for replacement of fresh media to the cells. Fresh media (12 mL) was added into the same T75 flask after collection of each supernatant (~12 mL of each) from the flask and on the respective day of collection (Figure 1). Before collecting the supernatants on each day, bright field images were taken for records. After collecting supernatants on the respective days post DENV2 infection, C6/36 culture supernatants were harvested by centrifugation at 1200 rpm (to pellet the cell debris/dead cells), and clear supernatants were aliquoted (from the respective timepoint) and stored at −80 °C in a freezer. After collecting the virus supernatants from DENV2-infected cells from the respective days, these viruses were used to perform the TCID50 assay to determine the viral infectivity. After determining the known viral titers from each of these timepoint supernatants, we infected naïve C6/36 mosquito cells and EA.hy926 endothelial cells (at 1 MOI for 72 h post infection) to determine the infection kinetics of these virus supernatants from different days in naïve recipient cells.

2.2. TCID50 Assay, Viral Infectivity Determination and Microscopy

To determine the infectious titer of these virus supernatants (collected from mosquito cell culture supernatants at days 3, 19, 33, 60, 90, 120, and 175), TCID50 (tissue culture infectious dose 50%) assay was performed as described before [15,17,40]. Briefly, 1 × 10⁴ C6/36 cells (in 180 µL of complete media) were seeded in clear-bottom 96-well black plates and allowed to adhere overnight. Following infection (with 20 µL of each dilution from every virus supernatant, creating a total of 200 µL per well) of naïve mosquito cells with these virus supernatants for 5 days, C6/36 cells were fixed with acetone–PBS mixture (3:1 ratio, for 20 min at −20 °C) and plates were air dried, washed with 1 × PBS and blocked with 5% FBS-PBS in 0.05% sodium azide solution for 15 min at RT. Each group of virus supernatants had six different dilutions (from 1 to 6) and at least 6–8 independent replicates, in addition to the uninfected negative controls. DENV2 viral E-protein was detected by incubation with 4G2 mouse monoclonal antibody (overnight at 4 °C) (catalog number NR-50327, obtained from BEI resources; Manassas, VA, USA) followed by three washes with 1 × PBS. Samples were incubated with Alexa-594 labeled mouse secondary antibody for 1 hour at RT, followed by washes (3X) with 1 × PBS. Image acquisition was performed for each dilution of every single virus supernatant with 6 replicates and by using the Cytation 7 imaging system (BioTek, Agilent, Santa Clara, CA, USA). Images were acquired as bright field (to show cells) and fluorescent (to detect viral E-protein through red channel and Alexa-594 staining) and these images (bright field and fluorescence) were overlaid to merge them. After imaging, C6/36 cells were scored based on the fluorescence and presence or absence of infection in comparison to the infected positive controls or uninfected negative controls (as internal controls). After determining the 50% tissue culture infectious dose (TCID₅₀), viral titers were determined as described before [15,17,40,41]. TCID50 is the amount of virus required to infect 50% of cells in a culture well. TCID50 values were converted to plaque-forming units (pfu)/mL to obtain the virus titers. Representative images from various dilutions at different days are shown. Images were obtained at 20× magnification. A scale bar (200 µm) is shown on each representative image from the respective groups. For bright field microscopy from T75 flask of C6/36 cells, images were acquired using a Samsung Galaxy S21 ultra smartphone camera (Samsung, Republic of Korea, with 108 MP, 12,000 × 9000 pixel resolution, and f/1.8 wide-angle lens or 48 MP, 8000 × 6000 pixels, and f/1.8 wide-angle lens with 1/1.33″ sensor). Scale calibration was performed using ImageJ2 software (NIH, Bethesda, MD, USA), where the software automatically detected the pixel-to-micrometer ratio from the reference measurement and generated the scale bar accordingly.

Figure 1. Schematic representation of virus supernatant collection and processing. C6/36 mosquito cells were infected with unknown titers of DENV2 to collect the virus supernatants at different timepoints of day 0 for uninfected samples (before infection and referred to as uninfected), 3, 19, 33, 60, 90, 120 and 175 post DENV2 infection. Fresh media was added after every viral supernatant collection to continue the viral replication process in host mosquito cells. On the day of collection, half of the virus supernatant was separated for molecular analysis (to perform TCID50 viral dilution assay, viral RNA extraction, cDNA synthesis and QRT-PCR analysis) and the other half of the virus supernatant was processed for EV isolation followed by QRT-PCR analysis to detect DENV2 loads.

Figure 1. Schematic representation of virus supernatant collection and processing. C6/36 mosquito cells were infected with unknown titers of DENV2 to collect the virus supernatants at different timepoints of day 0 for uninfected samples (before infection and referred to as uninfected), 3, 19, 33, 60, 90, 120 and 175 post DENV2 infection. Fresh media was added after every viral supernatant collection to continue the viral replication process in host mosquito cells. On the day of collection, half of the virus supernatant was separated for molecular analysis (to perform TCID50 viral dilution assay, viral RNA extraction, cDNA synthesis and QRT-PCR analysis) and the other half of the virus supernatant was processed for EV isolation followed by QRT-PCR analysis to detect DENV2 loads.

2.3. Isolation of Extracellular Vesicles (EVs) from Cell Culture Supernatants, EV Quantification and EV-Mediated Infection

EVs were isolated from half of the mosquito cell culture supernatants (6 mL) collected from each timepoint (days 3, 19, 33, 60, 90, 120, and 175) by the differential ultracentrifugation method and as described in our published studies [15,17,42,43]. Detailed procedures for EV isolations have been schematically shown in our previous studies [15,17]. EVs were quantified as described [41]. Briefly, isolated EVs from C6/36 cell culture supernatants were diluted (1:500) in 1% Tween-20 solution which was prepared in filtered 1 × PBS. We loaded 5 µL of the diluted samples onto nCS1 microfluidic cartridge TS-400 (obtained from the company Spectradyne Particle Analysis, Signal Hill, CA, USA) with a size range of 65–400 nm (in diameters). Loaded cartridges were inserted into an nCS1 (light-scattering particle) analyzer to quantify EVs from each sample. Data was collected and processed as described [41]. After collecting EVs from the indicated days, EA.hy926 cells and C6/36 naïve recipient cells were seeded in 12-well plates and infected via EVs (collected from 6 mL of previously collected cell culture supernatants) with 20 µL per well (6 replicates for each group in either of the naïve recipient cells) to determine the infection kinetics in these mosquito and endothelial cells.

2.4. RNA Extraction, cDNA Synthesis, QRT-PCR, Sequencing Analysis, and Gel Electrophoresis

Viral RNAs were extracted from DENV2-infected mosquito cell culture supernatants collected at days 3, 19, 33, 60, 90, 120, and 175 by using QIAamp Viral RNA Mini kits and the protocol recommended by the company (Qiagen, Germantown, MD, USA). Viral RNA was converted to cDNA using the iScript cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). The viral capsid transcript measurements are relative to the total viral RNA mass and not an absolute quantification of viral titers per volume or cell number. PCR reactions were performed with conditions of 95 °C for 2 min and with 39 cycles of 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 1 min and followed by 72 °C for 5 min. Lastly, reactions were then held at 4 °C. For determination of viral loads in naïve C6/36 cells and human endothelial cells, we incubated these cells with previously collected DENV2-infected C6/36 cell culture supernatants (from early timepoints of 3, 19, and 33 or from later timepoints of 60, 90, 120 and 175 days) at 1 multiplication of infection (MOI). Undiluted viral supernatants (with known viral titers determined by viral infectivity assay) were used for infection with 1 MOI, and the supernatant volumes for all timepoints of the virus were between 5 and 250 µL based on their titers. We seeded 4 × 10⁵ naïve C6/36 or EA.hy926 human endothelial cells in respective 12-well plates and prepared total RNA from respective cell lysates after 72 h co-incubation with viral supernatants. RNA was converted to cDNA, and we performed QRT-PCR and normalized to total cellular RNA. QRT-PCR reactions were performed with conditions of 95 °C for 3 min and with 39 cycles of 95 °C for 15 s, 58 °C for 30 s, and 72 °C for 40 s and followed by 95 °C for 10 s, 65 °C for 5 s and 95 °C for another 5 s. Lastly, reactions were then held at 4 °C. Upon DENV2 infection in mosquito/mammalian cells, some of the housekeeping genes like actin showed variations and hence we preferred using equal amounts of total RNA for quantification. Samples were collected on different days in Aurum total RNA lysis buffer (Bio-Rad, Hercules, CA, USA).

Total RNA was freshly isolated and quantified, and an equal amount of RNA was used to generate cDNA which is used as a template for amplification of the DENV2 capsid gene by quantitative real-time PCR (QRT-PCR) analysis. Equal amounts of cDNA samples were used for DENV2 capsid gene detection and QRT-PCR analysis. A CFX OPUS instrument (BioRad, Hercules, CA, USA) was used with conditions and oligonucleotides as described in our previous studies [15,17,44]. Data obtained from QRT-PCR were analyzed using the standard curve method, in which 10-fold serial dilutions were used to establish standards ranging from 1 to 1 × 10⁵ ng, where actual gene quantities were determined relative to different known standard concentrations as described in [40]. Samples were run in duplicate and with 5–6 replicates for QRT-PCR analysis. In all QRT-PCR runs, NTC (no-template controls) were included as negative controls. Viral loads and capsid gene transcripts were normalized to total viral RNA concentrations. Amplified QRT-PCR products were loaded into 1% agarose gel to perform gel electrophoresis. For sequence analysis, entire fusion loop regions of the Envelope gene and the entire NS5 gene were amplified by QRT-PCR and sequenced for identifying functional mutations. Fusion loop regions from all viral supernatants were amplified using the forward primer 5′ GCTGACCAACACAACAACAGATTCT 3′ and reverse primer 5′ CTGTGTCATTTCCGACTGCATGCT 3′. For amplification and sequencing of the full-length NS5 gene from the day 120 virus supernatant, the following primers were used and in this order: forward 5′ GGCTAACATTTTAGAGGGAGTTACT 3′ and reverse 5′ CCGCTTCCGAGGTCTACATCT 3′; forward 5′ CCGGGAACATAGTGTCATCAG 3′ and reverse 5′ CCATATGGCTCTGCTGCCTT 3′; forward 5′ GAGAACAAGTGGAAGTCGGCA 3′ and reverse 5′ CTGTCTGATTAGTTGGGCTTC 3′; forward 5′ GCGTGTGCAAAGACCAACACCA 3′ and reverse 5′ GTCACGTCTGTGGAAGTACATC 3′; and forward 5′ AATCATGAAAAGACGGCCGCG 3′ and reverse 5′ GGCATTTGTAATGGCCTGACTTCT 3′. The oligonucleotides used for amplification of DENV2 capsid gene transcripts are from published study [17].

2.5. Immunoblotting

Immunoblotting analysis was performed by following the published methods described in [15,17,44]. Briefly, 30 µL of virus supernatants (DENV2-infected mosquito culture supernatants collected at days 3, 19, 33, 60, 90, 120, and 175) was precleared with protein A/G beads (obtained from Peirce/Thermo/Fisher Scientific, Waltham, MA, USA) for reducing the serum proteins (from FBS) and then resolved on reducing condition with 12% SDS-PAGE gel. To visualize the total protein profile, 2,2,2-Trichloroethanol (TCE, obtained from Sigma-Aldrich, St. Louis, MO, USA) was incorporated into the gel mixture during preparation of the SDS-PAGE gel. After gel electrophoresis, total proteins were transferred onto nitrocellulose membrane by wet blotting overnight at 4 °C. Blots were blocked in 5% skim milk (obtained from Sigma-Aldrich, St. Louis, MO, USA) blocking buffer prepared in 1 × TBST buffer overnight at 4 °C. After blocking, blots were probed with respective primary antibodies for overnight incubation at 4 °C with gentle agitation/shaking. To detect DENV2 capsid protein, we used rabbit polyclonal antibody (catalog number ab155042, obtained from Abcam, Cambridge, MA, USA) in 1:1000 dilution or respective goat anti-rabbit secondary antibody with HRP conjugate (catalog number BA1054, obtained from BosterBio, Pleasanton, CA, USA). For detection of DENV2 membrane (M)-protein and prM protein, we used anti-mouse monoclonal dengue virus membrane (M) protein antibody or anti-mouse monoclonal dengue virus pre-membrane (prM) protein antibodies (catalog numbers ARG66680 and ARG66679, obtained from Arigo Biolaboratories, Zhubei city, Taiwan) and each in 1:1000 dilution. Respective secondary HRP-conjugated antibodies, goat anti-mouse IgG2b and goat anti-mouse Kappa (catalog numbers 1091-05 and 1050-05, obtained from Southern Biotech, Birmingham, AL, USA) were used to detect the dengue M or prM proteins, respectively. The reference band highlighted with an arrow (in total protein profile SDS-PAGE gel image) is pointed out in the immunoblot analysis.

2.6. Statistical Analysis

All comparative results in our data sets were analyzed by GraphPad Prism6 software. For comparing two unpaired means, two-tailed t-tests were performed for the entire analysis. Error bars represent mean (+SD) values. p values of 0.05 were considered for significant differences.

3. Results

3.1. Long-Term or Persistent DENV2 Infection Is Perceived in Mosquito Cells

We determined DENV2 loads present in mosquito cell culture supernatants collected from days 3, 19, 33, 60, 90, 120, and 175 post infection. A schematic representation summarizes the methodology used in this study (Figure 1). To assess the susceptibility of C6/36 cells upon DENV2 infection, images were captured at the various timepoints of days 0, 3, 19, 33, 60, 90, 120, and 175 post infection. Images were obtained from the growing mosquito cells and just before the harvesting of the supernatants from the T75 flask on that respective day. Images showed the morphology of C6/36 mosquito cells with persistent and long-term DENV2 infection at different days post infection (Supplementary Figure S1). These images illustrated gradual cytopathic effects (CPEs) shown from day 60, which became notably pronounced from 90 to 175 days post DENV2 infection. We cannot exclude that the CPEs could be due to natural cell death, but fresh cell culture media was replaced upon every timepoint of virus supernatant collection. We also noted stretched/elongated group of cells upon day 60 post infection (Supplementary Figure S1). Remarkably, the emergence of new cell patches kept growing in these cultures even after the longer timepoints of days 90–120, which further indicated the high resistance of C6/36 cells upon DENV2 infection (Supplementary Figure S1). Given that these cells had a DENV2 infection and were incubated for such longer timepoints of 90–120 days, we believe that these could be newly divided cells. These data indicate that DENV2 is capable of long-term infection and has an ability to persistently infect the highly permissive mosquito cells.

3.2. Determination of DENV2 RNA in Viral Supernatants Collected from Mosquito Cells

QRT-PCR analysis showed consistent detection of DENV2 RNA in viral supernatants collected from different days post infection, with no significant differences observed in viral capsid transcript levels (Figure 2A). DENV2 RNA product amplified from QRT-PCR analysis is shown on an agarose gel image (Figure 2B). Immunoblotting analysis further corroborated these findings of DENV2 viral RNA detection in cell culture supernatants (Figure 2C). Similar levels of DENV2 capsid protein were detected at different days (Figure 2C). These results indicated consistent detection of DENV2 RNA and proteins at all tested days. A total protein profile SDS-PAGE gel image served as a loading control for this immunoblotting analysis. The intense band on top of this profile gel image represents the serum protein. A protein band below the serum protein is considered for loading consistency. Taken together, these results confirmed the presence of DENV2 viral transcripts and protein in the mosquito cell culture supernatants collected at different days post infection and thus indicating a persistent DENV2 infection in mosquito cells.

3.3. Infectivity of DENV2 Decreased over the Longer Time of Infection

Detection of DENV2 capsid transcripts and protein from days 3–175 post infection in virus supernatants suggested a persistent infection of this virus in C6/36 mosquito cells. Next, we addressed whether this highly persistent and long-term DENV2 infectivity is maintained or affected in mosquito cells. TCID50 assay was performed to determine the DENV2 infectivity titers of these virus supernatants (collected from days 3 to 175 post infection) and as per described in our published studies [15,17,40]. Briefly, the virus supernatant titers were determined by identifying the dilution at which less than 50% infectivity was observed among the replicates of each dilution. Notably, TCID50 assay data revealed a decrease in the infectivity and titers of viral supernatants (collected from days 33 to 175 post DENV2 infection) (Figure 3). As evidenced by the gradual reduction in viral infection at selected dilutions (10⁻⁸, 10⁻⁷, 10⁻⁶, 10⁻⁵, 10⁻⁴, and 10⁻³), we observed less than 50% tissue culture infectivity when these respective viral supernatants infected naïve mosquito cells (Figure 3A). We noted that on day 5 post DENV2 infection, the dilution of 10⁻³ showed 50% infection, and similar data was obtained from virus supernatant collected at day 175 post DENV2 infection (Figure 3A). In addition, TCID50 assay further confirmed persistent DENV2 viral replication and long-term infectivity over the extended duration of infection (Figure 3A). Merged images are shown with high resolution and for clarity (Supplemental Figure S2). Quantification of infectivity from viral supernatants further showed decreasing virus titers over the course of DENV2 infection (Figure 3B). The viral supernatants collected at days 3 and 175 showed reduced viral titers followed by day 120 (Figure 3B). Titers determined from days 33, 60 and 90 showed similar infectivity between each other whereas the day 19 titer revealed the highest infectivity with DENV2 (Figure 3). Since infection from the timepoints of days 33–90 showed a similar trend but virus supernatants from days 120 and 175 revealed lower viral infectivity, we amplified the DENV2 Envelope (E) gene fusion loop fragments of viral supernatants by QRT-PCR and sequenced and analyzed this region that is critical for viral entry. The DENV2 E-gene fusion loop region from these viral supernatants is amplified and shown on an agarose gel image for further clarification (Supplementary Figure S3A). We did not find any changes or mutations in this E-gene fusion loop region from the 33-, 60-, 90- and 120-day viral supernatants (Supplementary Figure S3). The chromatogram peaks clearly revealed no changes in the nucleotide base pairs from any of the tested timepoints (Supplementary Figure S3B). Chromatograms showed clear, well-resolved peaks without any overlap, thus indicating high-quality sequencing data (Supplementary Figure S3B). In addition, we sequenced and analyzed the entire DENV2 NS5 gene (as part of the viral replication complex) and found two non-sense mutations at the nucleotide level, but these mutations revealed no amino acid substitutions (Supplementary Figure S4). The entire NS5 gene is shown as a schematic that represents the nucleotide mutations observed in the day 120 virus supernatant in comparison to the wild-type laboratory stock virus collected at day 19 (Supplementary Figure S4A). The sequence is shown along with the chromatogram peaks for each nucleotide base pair (Supplementary Figure S4B,C). Two nucleotide substitutions were identified at positions 2080 and 2203 in the day 120 viral supernatant (Supplementary Figure S4B). The nucleotide sequences of the mutated region show the corresponding sequencing chromatogram peaks confirming these substitutions (Supplementary Figure S4C). However, these nucleotide changes were synonymous and did not result in any alteration of the encoded amino acid sequence (Supplementary Figure S4B,C). These data showed that even though the DENV2 infection is longer and persistent, the viral infectivity and dynamics of infection are reduced over the long-term DENV2 infection in mosquito cells.

3.4. Infection Kinetics of DENV2-Infected Viral Supernatants Showed Reduced Viral Replication in Naïve Recipient Cells

To investigate the infection kinetics and viral replication efficiency (in naïve recipient cells) of viral supernatants (collected from days 3, 19, 33, 60, 90, 120 and 175 post DENV2 infection), both C6/36 cells and endothelial cells were infected (at 1 MOI for 72 h post infection) with respective viral supernatants. Remarkably, we found that DENV2 loads measured by QRT-PCR analysis showed an increasing trend in naïve recipient mosquito cells incubated with virus supernatants from days 3 to 19; however, we noted a significant decrease in infection loads with virus supernatants from the day 33 to 175 timepoints (Figure 4A). Like the reducing DENV2 infectivity revealed by TCID50 assay, viral supernatants from later timepoints also showed significantly decreased replication in naïve recipient mosquito cells. Furthermore, QRT-PCR analysis revealed that viral loads in naïve recipient endothelial cells had an increasing trend upon DENV2 infection with supernatants from days 19–33 that remained stable up to day 90 (Figure 4B). However, DENV2 loads were reduced upon infection with supernatants from days 120 to 175 (Figure 4B). Although endothelial cells showed stable DENV2 replication (upon days 19–90), both recipient cells had similar patterns of reduced viral loads upon infection with viral supernatants collected on days 120 and 175 post infection (Figure 4A,B). These results indicate a reduction in infection ability when DENV2 is propagated for a longer time in mosquito cells.

3.5. Infectious EVs from Viral Supernatants Revealed Reduced Infection in Naïve Recipient Cells

Our previous study reported the presence of a full-length DENV2 viral RNA genome inside EVs derived from mosquito cells [17]. Also, reduced viral infection and replication noted with cell culture supernatants (collected from days 33–175 post DENV2 infection) in mosquito and human endothelial naïve recipient cells further prompted us to investigate the effects of EVs collected from these supernatants. First, we isolated EVs from uninfected and DENV2-infected C6/36 cell culture supernatants (6 mL for each sample) collected from all timepoints of days 3–175. EVs were next quantified to determine their numbers and concentrations, as described in the Methods Section. We noted that EVs collected from days 0 (uninfected), 3, and 19 had lower EV numbers and concentrations when compared to the EV numbers and concentrations from the later timepoints of days 33, 60, 90, 120 and 175 post DENV2 infection (Figure 5A,B). The machine generated graphical data for EV concentrations and sizes for uninfected C6/36 cells for days 3, 19 and 33 (Supplementary Figure S5) and for days 60, 90, 120 and 175 is shown (Supplementary Figure S6). To determine the infection kinetics of EVs, we tested the early timepoints of days 3, 19 and 33 and the persistently infectious supernatants from days 33, 60, 90, and 120 post DENV2 infection. EVs collected from these supernatants (from days 3, 19, 33, 60, 90, and 120 post DENV2 infection and the uninfected control) were co-incubated (20 µL of EV suspension per well) on naïve C6/36 mosquito cells and endothelial recipient cells for 72 h. QRT-PCR analysis revealed a somewhat similar pattern observed between infection via cell culture supernatants in mosquito cells and that of treatment with infectious EVs collected from supernatants of days 3, 19, 33, 60, 90 and 120 post DENV2 infection (Figure 6). Infectious EVs collected from day 3 showed lower viral loads and at day 19 or 33, we detected higher DENV2 loads (Figure 6A). Also, naïve mosquito cells incubated with EVs from days 60–120 showed decreasing viral loads when compared to naïve C6/36 cells incubated with EVs from day 33 (Figure 6B). DENV2 infectious EVs from mosquito cells incubated on human endothelial cells showed reduced viral loads with the day 3 EVs, whereas EVs from days 19, 33 and 60 showed increased viral burden (Figure 6C,D). The viral loads in endothelial cells were significantly lower with EVs from days 90 and 120 when compared to the EV incubation with day 33 supernatants (Figure 6D). These data revealed that infectious EVs obtained from mosquito cell culture supernatants (collected from days 33–120 post infection) had reduced infection in naïve mosquito and endothelial recipient cells.

3.6. DENV2 Membrane Protein Is Present Only in EVs from Early Days

Immunoblotting analysis performed with EV lysates (collected from 3, 19, 33, 60, 90, 120 and 175 days post infection) detected membrane (M) protein only in EVs collected from days 3, 19 and 33 (Figure 7). We did not find any signals in the later timepoints (Figure 7). Also, no prM protein detection was observed in any of the tested EV lysates from any timepoints (Figure 7). It is noteworthy that M protein was present at higher levels in the day 33 EV lysates when compared to the day 3 and 19 lysates (Figure 7). Collectively, these findings indicated that although DENV2 persistently infected and survived for a longer time in mosquito host cells, its infectivity and replication efficiency is drastically reduced with this long-term and persistent infection.

4. Discussion

In this study, using C6/36 mosquito cells, we employed a long-term DENV2 infection model to explore the potential for its persistent infection within mosquitoes in nature. Numerous studies have highlighted the capability of arboviruses to establish prolonged infections in the vector host [18]. Vertical transmission in the vector host, wherein infected female mosquitoes can pass the virus to their offspring, represents one mechanism of facilitating DENV persistence between epidemics [18,23,36,37]. Investigating such infections in mosquitoes is challenging even under controlled conditions allowing survival for up to several days. Therefore, it is reasonable to use C6/36 Aedes albopictus mosquito cells as a model for studying persistent DENV infections [18,20]. Prior research has demonstrated that persistently DENV2-infected C6/36 cells produce virions (at 42 weeks or 294 days post infection) incapable of infecting BHK-21 baby hamster cells, with released viral particles exhibiting infectivity solely towards mosquitoes rather than vertebrate cells [18,20]. This adaptation may constrain viral fitness in the vertebrate cells [18]. Our findings revealed persistent DENV2 infection in C6/36 cells for up to 175 days. One of the limitations to this study is that the mosquito cells were seeded and infected with unknown numbers and viral titers, and cell culture supernatants were collected randomly without any bias but depending on the need to feed mosquito cells for longevity. Notably, each viral supernatant (collected on days 3, 19, 33, 60, 90, 120 and 175 post DENV2 infection in mosquito cells) retained the ability to infect both mosquito and human endothelial cells. However, infection kinetics data indicated significantly reduced viral loads and lower infectivity potential of viral supernatants from days 120 to 175 in both mosquito and human endothelial cells. We believe that due to the presence of high virus titers (as determined by the infectivity assay), an MOI of 1 is sufficient to infect the naïve recipient mosquito and endothelial cells with incubation for 72 h post infection. Lower viral titers determined at days 120 and 175 revealed lower viral replication in mosquito and endothelial cells. Both our study and the published report [18] consistently showed that DENV2 persistently infects mosquito cells but over a long course of infection, these viruses become deficient in viral infectivity and the replication process, especially in mammalian cells. Additionally, our results demonstrated that extracellular vesicles (EVs) derived from these viral supernatants at different timepoints mediate infection of mosquito and human endothelial cells in similar patterns. Collectively, these observations underscore the capacity of DENV2 for long-term persistent infection. Over the long-term persistent infection, we noted that C6/36 mosquito cells started exhibiting longer fiber-like structures that formed a meshwork-like appearance at day 33. Later in the course of infection at day 60, we found that cells appeared highly polarized and formed clusters of meshwork-like structures. We believe that these structures perhaps facilitate the dissemination of DENV2 from one mosquito cell to another. At 60 days post DENV2 infection, we started noticing cell death and the appearance of clumps of dead cells, which increased over the timepoint of day 90. By day 120 post DENV2 infection, sac-like structures appeared but were surrounded by healthy polarized cells. We also noted new colonies of mosquito cells near the sac-like structure. After day 175, we failed to keep these cultures as many of the cells appeared unhealthy and died.

It was noteworthy that DENV2 loads in virus supernatants were detectable from days 3–175 and at both the transcript and protein levels. The reduction in DENV2 viral infectivity and replication revealed that the longer incubation of mosquito cells reduces the virus’s ability to infect naïve host recipient cells. We found no differences between the mosquito and human endothelial cells as DENV2 loads reduced over the course of infection with virus supernatants from days 19–120 post infection. A similar infection pattern noted with EV-mediated infection further suggested that viral RNA/proteins and full-length viral RNA genomes packaged inside EVs are affected in virus supernatants, which resulted in reduced DENV2 loads. Overall, these data suggested that DENV2 can persistently infect mosquito cells for a longer term and perhaps maintain strong vector resistance and adaptability to arbovirus infection in nature. However, the viral infectivity and replication ability were compromised in both naïve mosquito vector cells and in naïve endothelial human recipient cells. Our previous work revealed increased viral loads in cells upon EV-mediated infection that had lower titers compared to direct infection via laboratory virus stocks with known higher titers. However, EVs collected from supernatants or virus supernatants at different days post DENV2 infection showed lower infection in naïve recipient cells. In summary, this study not only revealed the pathogenesis of DENV2 infection at different days post infection but also indicated compromised infectivity and replication of this virus when propagated for a long time in mosquito cells.