Autophagy Facilitates Antibody-Enhanced Dengue Virus Infection in Human Pre-Basophil/Mast Cells

Background Dengue virus (DENV) infection can cause severe hemorrhagic disease in humans. Although the pathogenic mechanisms underlying severe DENV disease remain unclear, one of the possible contributing factors is antibody-dependent enhancement (ADE) which occurs when sub-neutralizing antibodies derived from a previous DENV infection enhance viral infection through interaction between virus-antibody complexes and FcR-bearing cells, such as macrophages and basophil/mast cells. Although recent reports showed that DENV induces autophagy, the relationship between antibody-enhanced DENV infection and autophagy is not clear. Methodology/Principal Findings We showed that sub-neutralizing antibodies derived from dengue patient sera enhanced DENV infection and autophagy in the KU812 pre-basophil-like cell line as well as the HMC-1 immature mast cell line. Antibody-enhanced DENV infection of KU812 cells increased the number of autophagosome vesicles, LC3 punctation, LC3-II accumulation, and p62 degradation over that seen in cells infected with DENV alone. The percentages of DENV envelope (E) protein-positive cells and LC3 puncta following antibody-enhanced DENV infection of KU812 cells were reduced by the autophagy inhibitor 3-MA. Antibody-enhanced DENV infection of HMC-1 cells showed co-localization of DENV E protein and dsRNA with autophagosomes, which was inhibited by 3-MA treatment. Furthermore, DENV infection and replication were reduced when KU812 cells were transfected with the autophagy-inhibiting Atg4BC74A mutant. Conclusions/Significance Our results demonstrate a significant induction of autophagy in antibody-enhanced DENV infection of pre-basophil-like KU812 and immature mast cell-like HMC-1 cells. Also, autophagy plays an important role in DENV infection and replication in these cells. Given the importance of ADE and FcR-bearing cells such as monocytes, macrophages and basophil/mast cells in dengue disease, the results provide insights into dengue pathogenesis and therapeutic means of control.

Basophil/mast cells expressing Fc receptors play an important role in a wide variety of inflammatory reactions and in host defense against pathogens. These cells selectively produce and secrete numerous factors including chemokines, cytokines, lipid mediators, and granule-associated products [20]. In dengue patients, increased levels of blood and urinary histamine, a major granule product of basophil/mast cells, have been reported to correlate with disease severity [21,22]. A large histological study of dengue hemorrhagic fever (DHF) patients noted mast cell activation as demonstrated by swelling and vacuolation of the cytoplasm, and loss of granule integrity [23].
Autophagy is an evolutionarily ancient pathway which contributes to cell survival. Conditions such as nutrient starvation, pathogen infection, and other environmental stresses can induce autophagy. During autophagy, portions of the cytoplasm or small organelles are sequestered into double-membrane vesicles called autophagosomes. Autophagosomes ultimately fuse with lysosomes to generate single-membrane vesicles termed autophagolysosomes, where the contents are subsequently degraded. Transmission electron microscopy (TEM) has been used to detect autophagosome formation, and the expression level of LC3-II is another indicator [24]. Autophagy may play both anti-viral and pro-viral roles in viral infection and pathogenesis [25]. Previous studies showed that DENV induces autophagy in human hepatoma cell lines and peripheral blood monocytes to promote viral replication [26][27][28]. A DENV nonstructural protein, NS4a, has been reported to induce autophagy to protect cells against death and enhance DENV replication [29]. In contrast, the induction of autophagy has been reported to reduce DENV output in monocytic cell line [30]. It has also been shown that DENV triggers autophagy to regulate lipid metabolism [31,32]. A previous study by Ubol et al. showed that entry of DENV-antibody complexes into human monocytic cells activated negative regulators DAK and Atg5-Atg12, which then disrupted the RIG-I/MDA-5 signaling cascade and reduced type-1 interferon (IFN) production. This would lead to suppression of IFN-mediated antiviral responses [33]. Their study indicated the involvement of autophagy-related proteins Atg5-Atg12 to evade host innate immune responses. In the present study, we investigated the induction of autophagy by antibodyenhanced DENV infection of pre-basophil-like KU812 cells and immature mast cell-like HMC-1 cells. We provide evidence of a definitive link between antibody-enhanced DENV infection and autophagosome formation and elucidate mutual regulatory pathways between virus replication and induction of autophagy.

Dengue virus culture
Dengue virus serotype 2 (DENV2, strain 16681) originally isolated from a Thai patient who suffered from DHF [39] was used throughout this study and was maintained in C6/36 cells. Briefly, monolayers of C6/36 cells were incubated with DENV at a multiplicity of infection (MOI) of 0.01 and incubated at 28uC in 5% CO 2 for 5 days. The cultured medium was harvested and cell debris was removed by centrifugation at 9006g for 10 min. After further centrifugation at 16,0006g for 10 min, the virus supernatant was collected and stored at 280uC until use. Virus titer was determined by plaque assay using the BHK-21 cell line.

Dengue patient sera
For ADE assay of DENV infection, a dengue-immune serum pool was obtained from nine convalescent-phase sera from patients recovering from DENV2 infection. Dengue-convalescent patient sera were collected in Thailand in 1990 as part of long-standing surveillance and provided by Dr. Bruce Innis (Armed Forces Research Institute of Medical Science, Bangkok, Thailand) and described previously [40].

Dengue virus infection
Aliquots of DENV were resuspended with or without 1:10,000 dilution of pooled dengue patient sera for 1 h at 4uC. KU812 or HMC-1 cells were incubated with DENV (with or without pooled dengue patient sera) at MOI of 1 for 90 min at 4uC. Cells were then washed twice with RPMI medium to remove unabsorbed virus and antibodies. Cells were resuspended and supplemented with 2% FBS-containing medium at 37uC for further incubation.
Plaque assay BHK-21 cells were plated onto 12-well plates (1610 5 cells/well) and cultured in DMEM under CO 2 -enriched conditions. Supernatants and cell lysates from DENV-infected cells were serially diluted and inoculated with BHK-21 cells for plaque assay. After 2 h post-infection, the solution was replaced with fresh DMEM containing 2% FBS and 0.5% methyl cellulose (Sigma-Aldrich). At five days post-infection, the medium was removed, and the cells were fixed and stained with 1% crystal violet, 0.64% NaCl, and 2% formalin (Sigma-Aldrich).

Flow cytometry analysis
Following DENV infection, cells were washed with PBS, fixed with 1% formaldehyde, and permeabilized with 0.1% saponin (Sigma-Aldrich) at room temperature for 10 min. Fc receptors of cells were blocked with 1:100 dilution (in permeabilizing buffer) of normal human sera (approved by the Institutional Review Board of National Cheng Kung University Hospital, No. A-ER-102-123) at 4uC for 1 h. After washing, cells were then stained with anti-DENV envelope (E) protein or anti-nonstructural protein 4B (NS4B) (GeneTex) at 4uC for 30 min. Cells were incubated with Alexa488-conjugated secondary antibody (Life Technologies) at 4uC for 30 min and analyzed using FACS Calibur (BD Biosciences). For the anti-E antibody-enhanced DENV infection experiment, cells were then stained with FITC-conjugated anti-E antibodies at 4uC for 1 h and analyzed using FACS Calibur. For the Atg4B mutant-transfected antibody-enhanced DENV infection experiment, cells were stained with anti-NS4B antibodies at 4uC for 30 min, followed by Alexa647-conjugated secondary antibody (Life Technologies) at 4uC for 30 min, and analyzed using an LSRFortessa instrument (BD Biosciences).

Immunofluorescence
Cells were fixed with 1% formaldehyde (Sigma-Aldrich), permeabilized with 0.1% saponin, and then blocked Fc receptors with normal human sera. Cells were then stained with anti-E, antidouble strand RNA (dsRNA) (English & Scientific Consulting), or anti-LC3 (MBL) antibodies at 4uC for overnight. Cells were then incubated with Alexa488-or Alexa594-conjugated secondary antibody (Life Technologies) at 4uC for 1 h and analyzed with a FV1000 confocal microscope (Olympus).

Cell transfection
The plasmids (mStrawberry and mStrawberry-Atg4B C74A ) were gifts of Dr. Tamotsu Yoshimori (Osaka University, Japan). Briefly, the Atg4B cDNA was cloned from genomic DNA isolated from mouse embryonic fibroblast (MEF) cells and was inserted into pmStrawberry-C1 using engineered BamHI and KpnI sites; the point mutation was introduced using the QuikChange Site-Directed mutagenesis system (Stratagene) [41]. For cell transfection, KU812 cells (1610 7 cells) were resuspended in 0.6 ml RPMI-1640 medium containing 10 mg/ml BSA, and mixed with 20 mg plasmid DNA in electroporation cuvette (0.4-cm electrode gap) and subjected to 220 V for 70 msec by Gene Pulser II Electroporation System (Bio-Rad). After electroporation, the cells were washed with fresh medium and then cultured for 48 h before DENV infection.

Transmission electron microscopy
Twenty-four h after DENV infection, KU812 cells were fixed with 2.5% glutaraldehyde (Sigma-Aldrich) in 0.1 M cacodylate buffer containing 4% sucrose, 1 mM MgCl 2 and 1 mM CaCl 2 , and post-fixed in 1% osmium tetroxide (Sigma-Aldrich). The cells were further dehydrated with ethanol and embedded with LR White. Ultrathin sections were stained with uranyl acetate and lead citrate (Sigma-Aldrich), and then observed using a HT7000 transmission electron microscope (Hitachi).

Western blot analysis
Following infection with DENV, cells were lysed using a Triton X-100 based lysis buffer with a protease inhibitor mix and phosphatase inhibitors (Sigma-Aldrich). Cell extract was separated using SDS-PAGE and transferred to polyvinylidene difluoride membrane (Millipore). After blocking with 5% nonfat milk in PBS-T (0.05% Tween-20), blots were probed with anti-LC3, -NS4B, -p62 (MBL), and -b-actin (Santa Cruz Biotechnology) antibodies at 4uC for overnight. After washing with PBS-T, blots were treated with horseradish peroxidase (HRP)-conjugated secondary antibodies (Cell Signaling Technology) in 1:5000 dilution at room temperature for 1 h. Blots were developed using western lightning chemiluminescence reagent (Perkin Elmer for all blots, except Millipore for LC3 blot).

Statistics
We used the paired t-test for statistical analysis. Statistical significance was set at P,0.05.

Antibody-enhanced DENV infection and autophagosome formation in KU812 cells
Recent reports demonstrated that autophagy is observed in several kinds of DENV-infected cells [42]. However, the relationship between antibody-enhanced DENV infection and autophagy is still unclear. Human pre-basophil-like KU812 cells were inoculated with DENV (MOI = 1) in the presence or absence of dengue patient sera (1:10,000 dilution). Detection of DENV infection in KU812 cells was performed by flow cytometry ( Figure 1A and 1B) and plaque assay ( Figure 1C and 1D). Our previous study indicated very few virus-positive cells when KU812 cells were inoculated with DENV in combination with normal (i.e. dengue non-immune) human sera [43]. In addition, although the possibility of non-specific effect of antibody-containing serum to other RNA viruses cannot be completely ruled out, it was previously reported that infection of KU812 cells with another virus (RSV) was not enhanced by human RSV-positive, denguenegative sera [43]. We herein showed the intracellular expression of E protein and NS4B protein at 24 h post-infection was significantly enhanced by antibody-enhanced DENV infection ( Figure 1A and 1B). In addition, virus titers from the supernatant of antibody-enhanced DENV-infected KU812 cells ( Figure 1C) and the combined supernatant and infected cells ( Figure 1D) were increased than those from KU812 cells infected with DENV alone. UV-inactivated DENV (iDENV) served as negative control ( Figure 1A-1D). Furthermore, we found that antibody-enhanced DENV infection of KU812 cells increased autophagosome formation as observed by electron microscopy ( Figure 1E, arrowhead). The number of autophagosomes from one section in each cell was quantified from Figure 1E and shown in Figure 1F.
We further confirmed autophagy in DENV-infected KU812 cells using confocal microscopy to analyze the expression of DENV E protein and LC3 punctation, as a marker for autophagy. In addition to increased DENV E protein expression (Figure 2A, filled arrowhead), antibody-enhanced DENV infection also increased LC3 punctation (Figure 2A, empty arrowhead). Furthermore, KU812 cells were infected with iDENV in the presence and absence of subneutralizing dengue patient sera. The LC3 punctation also apparently increased in the iDENV infection of KUB12 cells with subneutralizing dengue patient sera ( Figure 2A). The co-localization of E-protein and LC3 punctation is also shown (merge, zoom, yellow). The quantified results of E protein expression and LC3 punctation from Figure 2A are shown in Figure 2B and Figure 2C, respectively. Not every cell expressing E protein showed LC3 punctation and vice versa. We therefore further analyzed the percentages of cells with both E protein expression and LC3 punctation (Figure 2A, merge, arrow, and Figure 2D). Cell lysates collected from cells infected with DENV alone or antibody-enhanced DENV were analyzed by Western blotting to confirm autophagy induction and viral infection. LC3-II accumulation and p62 degradation as autophagy indicators as well as DENV NS4B expression were increased in the antibodyenhanced DENV-infected cells ( Figure 2E). In addition, nutrientrich medium incubation served as a negative control and stimulation of autophagy by starvation served as a positive control of autophagy ( Figure 2E). It has been previously demonstrated that dsRNA, as an indicator of DENV replication, can co-localize with LC3 punctation structures in hepatocytes [28]. Here, we found that the number of dsRNA-positive KU812 cells ( Figure S1, filled arrowhead) as well as LC3 puncta ( Figure S1, empty arrowhead) were increased in antibody-enhanced, compared to DENV alone, infection. Also, numbers of dsRNA/LC3 punctation-positive KU812 cells were increased in antibody-enhanced DENV infection ( Figure S1, merge, arrow). The co-localization of dsRNA and LC3 punctation is also shown ( Figure S1, zoom, yellow). It is noteworthy that co-localization of dsRNA and LC3 punctation occurs sporadically, indicating either a transient interaction or an interaction involving only a subset of the total dsRNA. Furthermore, the autophagy inhibitor 3-MA was used to assess if modulation of autophagy may alter DENV infection. KU812 cells were infected with DENV alone or in combination with enhancing antibody and with or without 3-MA. Results showed that treatment with 3-MA inhibited DENV E and NS4B protein expression in KU812 cells whether they were inoculated with DENV alone or with enhancing antibodies (Figure 2F and 2G). A representative histogram of each group is shown in Figure S2.
We further used purified anti-DENV E protein monoclonal antibody 137-22 (a gift from Dr. Huan-Yao Lei's laboratory) [44] at a sub-neutralizing level, and results showed that the percentages of E-positive cells were significantly enhanced by DENV infection in the presence of 0.1 mg/ml purified anti-E monoclonal antibody over that seen in cells infected with DENV alone ( Figure 3A). Furthermore, the percentages of E-positive cells were reduced by 3-MA treatment following DENV infection with or without subneutralizing anti-E antibody ( Figure 3A). The confocal microscopy results also showed anti-E antibody-enhanced DENV infection ( Figure 3B, filled arrowhead). Moreover, LC3 punctation was induced by DENV infection in the presence of sub-neutralizing anti-E antibody ( Figure 3B, empty arrowhead). Some cells expressed both E protein and LC3 punctation ( Figure 3B, arrow). LC3 punctation was also reduced by 3-MA treatment ( Figure 3B). The quantified results are shown in Figure 3C.  Figure 4A). Co-localization of E proteins with LC3 punctation was also observed in some (dotted arrow), but not all, HMC-1 cells. When HMC-1 cells were treated with autophagy inhibitor 3-MA, the expression of E and LC3 punctation and their co-localization were inhibited ( Figure 4A). Similar findings were observed in the expression of dsRNA and its co-localization with LC3 punctation following antibody-enhanced DENV infection of HMC-1 cells ( Figure 4B). The enhanced expression of dsRNA and its co-localization with LC3 punctation were also inhibited when cells were treated with 3-MA ( Figure 4B).

Autophagy facilitates DENV infection and antibody-enhanced DENV infection in KU812 cells
In order to further confirm the role of autophagy, we established the Atg4B C74A -expressing KU812 cells. Atg4B C74A is an inactive Atg4B mutant due to the mutation of the catalytic cysteine residue (Cys74) and in consequence lacks protease activity. Atg4B C74A can  hamper autophagosome closure and has been utilized as a useful tool to inhibit autophagy [45]. Cells transfected with empty LC3, p62, and NS4B from total cell lysates were detected by Western blotting. b-actin served as internal control. NC: negative control (nutrient-rich medium); PC: positive control (starvation; Hank's balanced salt solution). (F and G) KU812 cells were pre-treated with or without 5 mM 3-MA for 1 h before incubation with medium alone (Mock), DENV alone, or DENV with sub-neutralizing dengue patient sera. 3-MA was maintained in the medium during DENV infection. After 24 h post-infection, the expression of DENV E protein (F) and NS4B protein (G) was detected by flow cytometry. The means 6 SD of three independent experiments are shown. **P,0.01, ***P,0.005. doi:10.1371/journal.pone.0110655.g002  strawberry plasmids served as control. We first confirmed that LC3 punctation was inhibited in mutant Atg4B C74A -expressing cells after transfection with Strawberry-Atg4B C74A plasmids in nutrient-rich and starvation conditions ( Figure S3). Since not all cells were successfully transfected with Strawberry (transfection efficiency 18.3%) or Strawberry-Atg4B C74A plasmids (transfection efficiency 14.1%), we analyzed the infectivity only from the cells showing red fluorescence. We counted the same numbers of Strawberry-and Strawberry-Atg4B C74A -expressing cells (red fluorescence; Figure 5A, arrowhead) to further determine the DENV-infected cells, which expressed DENV E protein (green fluorescence; Figure 5A, merge, arrow). Quantified results from Figure 5A showed that the percentage of E-positive cells was increased in antibody-enhanced strawberry-expressing cells (Figure 5B). Besides E protein, we also confirmed that the percentage of NS4B-positive cells was increased in antibody-enhanced strawberry-expressing cells by flow cytometry ( Figure 5C). Furthermore, in both strawberry-Atg4B C74A -expressing cells infected with DENV alone or with enhancing antibody, the percentages of E and NS4B-positive cells were inhibited ( Figure 5B and 5C). Reduced viral titers were also found in culture supernatants ( Figure 5D) and the combined supernatant and lysates from strawberry-Atg4B C74A -expressing cells ( Figure 5E). Since LC3 is also involved in the autophagosome formation [24], we also verified the role of LC3 in DENV infection. Cells were transfected with plasmid expressing shLuc (control) or shLC3. Successful knock-down of shLC3 expression was demonstrated by Western blot analysis ( Figure S4A). DENV infection of KU812 cells as determined by viral NS4B protein expression was also decreased when cells were transfected with shLC3 as compared with cells treated with control shLuc (Figure S4B and S4C). Taken together, the results indicate that inhibition of autophagy reduces DENV infection and replication in KU812 cells whether inoculated with DENV alone or with enhancing antibody.

Discussion
Previous reports have noted that DENV infection is associated with the autophagic response. DENV infection can induce ER stress [46], and ER stress has been demonstrated to activate autophagy [47]. In addition, DENV and antibody binds to FcR and activates DAK and Atg5-Atg12, which act as negative regulators to disrupt RIG-I/MDA-5 signaling and consequently type-1 IFN production [33,48]. Several pathogens, such as poliovirus and rhinovirus, employ autophagy to evade host immune responses [49]. DENV utilizes components of the cellular autophagy pathway for its own growth benefit [26][27][28]31]. A recent study using a novel autophagy inhibitor spautin-1 showed a direct effect of constituents of the autophagy pathway on DENV assembly [50]. Taken together, these findings suggest that pharmacological inhibition of autophagy as therapeutics against DENV infection and pathogenesis may be promising. In the present study, we showed that DENV replication and viral yield were significantly decreased when pre-basophil-like cells were treated with autophagy inhibitor 3-MA or overexpression of the mutant protein Atg4B C74A . DENV also utilizes autophagosomes for replication in human hepatoma cell lines [26][27][28]51]. While autophagy was induced in response to ADE-mediated DENV2 infection of monocytic U937 cells, biochemical induction of autophagy actually resulted in a decreased virus output and downregulation of autophagy caused only slight increase in DENV levels, suggesting that autophagy may not play a significant role of DENV replication in monocytic cells [30]. Therefore, there are probably cell-type differences in the importance of DENVautophagy interaction [30].
Although the autophagosome may be the site for DENV replication, our results showed that DENV E proteins and dsRNA co-localized with autophagosomes only in some DENV-infected pre-basophil-like cells. This finding suggests that autophagy may also indirectly promote DENV replication in these cells. Nevertheless, we cannot exclude the possibility of limited detection sensitivity in our immunofluorescence staining assay. It was previously shown that GFP-LC3 puncta did not co-localize with DENV replication markers [31]. In contrast, other studies showed the co-localization of DENV and autophagosomes, and this was observed at least 24 h post-infection [27,28,30]. Using a singlevirus particle tracking technology to determine whether DENV interacts with autophagy machinery, Chu et al. showed the colocalization of DENV and LC3 puncta. They found that viral particles interacted directly with autophagosomes at an early stage of DENV infection in a hepatoma cell line and promoted DENV replication [51].
Heaton et al. demonstrated that DENV-induced autophagic degradation modulates cellular lipid metabolism which indirectly promotes DENV replication in infected cells [52]. Because metabolism of lipid droplets is important for mast cell differentiation and activation to further release inflammatory mediators [53], it is worthwhile to further explore whether autophagic degradation of cellular lipid droplets is involved in the promotion of DENV replication. On the other hand, we also observed that some cells show autophagy-related LC3 punctuation despite the absence of detectable DENV infection. It has been demonstrated that inflammatory cytokines IFN-c or MIF can initiate an intracellular autophagic response [54][55][56]. Because the production of IFN-c and MIF is also associated with DENV disease, these and/or other DENV-associated cytokines may play a role in triggering the autophagy response in non-infected cells.
Ushio et al. suggested that autophagy plays a crucial role in degranulation of mast cells [57]. In this study, we have proposed a novel role for autophagy that may explain immune-mediated pathology by DENV. In dengue patients, serum and urinary levels of histamine, a major granule component of basophil/mast cells, were increased and correlated with disease severity [22]. DHF and dengue shock syndrome patients also show increased serum levels of anaphylatoxins C3a, C5a, and IgE, which can activate basophil/mast cells and contribute to vascular leakage possibly leading to shock [58]. In vitro studies also showed that DENVinfected basophil/mast cells can release potent vasoactive cytokines, such as IL-1b and IL-6, and chemokines, such as CCL3, CCL4, and CCL5 [15,16,43]. DENV-induced basophil/mast cell activation elevated systemic vasoactive products including leukotrines and chymase, which may contribute to vascular leakage during DENV infection [58]. Basophil/mast cells may also play a protective role in the host response to DENV because blocking these cells or their mediators has been reported to reduce DENV clearance [59,60]. The role of basophil/mast cells in DENV pathogenesis is still not fully defined and needs to be further investigated.
In DENV-endemic areas of the world, pre-existing subneutralizing antibodies may promote DENV infection via facilitating viral binding and entry into FcR-bearing cells by the mechanism of ADE [61,62]. Besides an extracellular or extrinsic role of enhancing antibodies, ADE may also enhance infection intrinsically by regulating the production of inflammatoryassociated cytokines, such as IL-6, IL-10, IL-12, TNF-a and IFN-c [63,64]. ADE may also disrupt innate pathways of anti-viral response [33,48]. In basophil/mast cells, the potent vasoactive cytokines, IL-1b and IL-6, and chemokines, CCL3, CCL4, and CCL5 are increasingly released after DENV infection with enhancing antibodies [15,16,43]. The release of vasoactive cytokines and chemokines causing vascular leakage on a systemic level can contribute to DENV pathogenesis [65].
So far, it is still unclear what mechanism regulates the induction of autophagy in ADE-mediated DENV infection. We found that autophagy was still induced when cells were infected with UVinactivated DENV in the presence of sub-neutralizing antibodies ( Fig. 2A and 2C). This result suggests that ligation of Fc receptor with complex of DENV and enhancing antibodies may activate Fc receptor signaling and trigger the autophagy response. Therefore, the role of Fc receptor signaling in inducing the autophagy response during antibody-enhanced DENV infection needs to be further explored.
In addition to autophagy, basophil/mast cells exposed to antibody-enhanced DENV infection also trigger apoptosis [18]. DENV infection alone may also trigger apoptosis in murine and human monocytes or neurons [66][67][68]. However, other reports showed that DENV-infected cells are protected from cell death. Furthermore, flavivirus NS4A-induced autophagy protects cells against cell death and enhances viral replication [29]. The mechanisms by which apoptosis and autophagy are regulated in antibody-enhanced DENV-infection are intriguing and require further investigation.
In conclusion, several mechanisms have been previously proposed for ADE in dengue disease pathogenesis including promoting viral entry into FcR-bearing cells, the induction of IL-10 to inhibit the IFN-mediated anti-viral pathway, and the activation of DAK and Atg5-Atg12 to inhibit IFN production [33,48]. In the present study, we demonstrate an additional mechanism in which autophagy is induced in antibody-enhanced DENV infection and leads to increased viral replication in prebasophil-like cells. Our finding clearly links autophagy and antibody-enhanced DENV infection. Given the importance of ADE and FcR-bearing cells in dengue disease, the results of this study may provide insights into dengue pathogenesis and therapeutic strategies. Figure S1 Co-localization of dsRNA with autophagosomes in KU812 cells after infection with DENV alone or with enhancing antibody. KU812 cells were incubated with medium alone (Mock), with DENV alone, or with DENV in the presence of sub-neutralizing dengue patient sera. After 24 h postinfection, cells were fixed, permeabilized, and stained with antidouble strand (ds) RNA (red), LC3 protein (green), and DAPI (blue). Cells were then mounted and observed by confocal microscopy. The filled arrowheads indicate the dsRNA-positive cells, the empty arrowheads indicate the cells with LC-3 punctation, and the arrows indicate the cells in which dsRNA is co-localized with LC3 punctation. The square areas are zoomedin images and shown in the right panels (merge, zoom). Bar: 10 mm. The imaging data were repeated three times and one set of representative results is shown. (TIF) Figure S2 The autophagy inhibitor 3-MA reduces DENV infection. KU812 cells were pre-treated with or without 5 mM 3-MA for 1 h before incubation with medium alone (Mock), DENV alone, or DENV with sub-neutralizing dengue patient sera.