Neutralizing Antibodies Inhibit HIV-1 Infection of Plasmacytoid Dendritic Cells by an FcγRIIa Independent Mechanism and Do Not Diminish Cytokines Production

Plasmacytoid dendritic cells (pDC) expressing FcγRIIa are antigen-presenting cells able to link innate and adaptive immunity and producing various cytokines and chemokines. Although highly restricted, they are able to replicate HIV-1. We determined the activity of anti-HIV-1 neutralizing antibodies (NAb) and non-neutralizing inhibitory antibodies (NNIAb) on the infection of primary pDC by HIV-1 primary isolates and analyzed cytokines and chemokines production. Neutralization assay was performed with primary pDC in the presence of serial antibodies (Ab) concentrations. In parallel, we measured the release of cytokines and chemokines by ELISA and CBA Flex assay. We found that NAb, but not NNIAb, inhibit HIV-1 replication in pDC. This inhibitory activity was lower than that detected for myeloid dendritic cells (mDC) infection and independent of FcγRIIa expressed on pDC. Despite the complete protection, IFN-α production was detected in the supernatant of pDC treated with NAb VRC01, 4E10, PGT121, 10-1074, 10E8, or polyclonal IgG44 but not with NAb b12. Production of MIP-1α, MIP-1β, IL-6, and TNF-α by pDC was also maintained in the presence of 4E10, b12 and VRC01. These findings suggest that pDC can be protected from HIV-1 infection by both NAb and IFN-α release triggered by the innate immune response during infection.

and not the other FccRs 25 . Moreover, pDC produce cytokines and chemokines, and in particular large amounts of type I interferon (IFN) in the presence of HIV 26 . These cells play an important role in innate immune response (reviewed in 27,28 ) and in the triggering of the adaptive immune response by presenting antigens, although with less efficiency than mDC (reviewed in 29,30 ). Mean of the CD83 MFI (mean of fluorescence intensity) measured in each experiment is represented for pDC treated under the same conditions described in (D). One-way ANOVA showed no significant difference between these conditions. Data are means 6 standard deviation (SD) of independent experiments performed with primary pDC from at least three different healthy donors. Groups were compared by one-way ANOVA (Kruskal-Wallis test), with P , .05 considered significant. In this study, we assessed the capacity of Ab (NAb and NNIAb) to inhibit HIV-1 replication in pDC, and analyzed the involvement of FccRIIa in this inhibition. In addition, the production of IFN-a, MIP-1a, MIP-1b, IL-6, and TNF-a by pDC was determined following inhibition of pDC infection by HIV-1 specific Ab.

Results
Purified Primary pDC Express FccRII and are Infected by HIV-1. Before analyzing the inhibitory activity of Ab and the role of FccRII in this inhibition, the expression of FccRII by living CD123-positive pDC and the infection of these cells were firstly evaluated by flow cytometry. We found that 41.2% of living CD123-positive pDC expressed FccRII after BDCA-4 positive selection (Fig. 1A). This result agrees with previous reports showing that 30 to 50% of FccRIIa, and not other FccRs, was detected on pDC present in the population isolated from blood circulation 25 . Thus, BDCA-4 positive selection was further used in this study to investigate the role of FccRIIa in the inhibition of HIV-1 replication by Ab. HIV-1 infection of pDC were detected by flow cytometry according to the size and the structure (Fig. 1B, left dot plot); the living population ( Fig. 1B, right dot plot) was selected and the percentage of p24 1 cells among the CD123 1 population was determined (Fig. 1C, left dot plot, mock condition). 3.5% of pDC were positive for p24 after 48 h of culture with HIV-1 BaL primary isolate, compared to only 0.1% of the AZT controls (Fig. 1C). Inhibition of HIV-1 BaL infection of the pDC was analyzed for a panel of anti-HIV-1 Ab by flow cytometry as described above. The viability of pDC was similar for the different healthy blood donors and for all experimental conditions, i.e.: approximately 70% of pDC were alive after 48 h of incubation with or without HIV-1 and with or without anti-HIV-1 specific Ab (Fig. 1D). Expression of the maturation marker CD83 at the surface of pDC varied depending on the donors but did not differ significantly after infection and in the presence of anti-HIV-1 Ab (Fig. 1E). Therefore, neither HIV-1 nor Ab influenced the viability and the maturation of the pDC. We further used these experimental conditions to study the inhibitory activity of anti-HIV-1 specific Ab and the role of FccRIIa in this inhibition.
Neutralizing Antibodies Inhibit HIV-1 Infection of Primary pDC. The inhibitory activity of NAb on pDC infection was analyzed. The percentage of HIV-1 BaL -infected pDC in the presence of NAb  compared to infected pDC in the absence of NAb (infection control) was represented for various concentrations of NAb 2F5 and VRC01 ( Fig. 2A). The concentration of Ab required to inhibit 90% of the infection (IC 90 ) was determined for each anti-HIV-1 Ab and the mean of the IC 90 from different healthy donors was calculated. The mean IC 90 of NAb 2F5 and VRC01 were of 80 mg/ml and 7 mg/ml, respectively (Fig. 2B). Moreover, the most recent broadly anti-gp120 NAb PGT121 and 10-1074 efficiently inhibited pDC infection by HIV-1 BaL with an IC 90 of 1.3 mg/ml and 6 mg/ml, respectively; anti-gp41 NAb 10E8 also inhibit pDC infection but with lower efficiency (Fig. 2C, D and E, and Table 1). Moreover, the inhibitory activity of NAb was confirmed on pDC infected with subtype C HIV-1 TV1 (e.g., IC 90 of 100 mg/ml for 2F5) and with transmitted/founder HIV-1 Bx11 viruses 31 (e.g., IC 90 of 3.5 mg/ml for PGT121). In parallel, we analyzed the inhibitory activity of these Ab on MoDC infection using similar experimental conditions as for pDC 17 . Interestingly, we observed a 7-80 fold lower IC 90 on MoDC (Table 1), indicating that pDC infection was more refractory to Ab neutralization than that of MoDC. We next measured the inhibitory activity of non-neutralizing HIV-1 specific Ab 246-D, 4B3 and 3D6. These Ab recognize the principal immunodominant domain (PID) of gp41. Anti-PID Ab 246-D and 4B3 are NNIAb and inhibited the infection of MoDC with an IC 90 of 45 mg/ml (Table 1). Despite the inhibitory activity on MoDC, these anti-PID Ab had no effect on pDC infection even at a concentration of 200 mg/ml ( Fig. 2F and Table 1). Ab 3D6 did not display any inhibitory activity on the infection of MoDC or pDC by HIV-1, and thus was referred to as a non-neutralizing non-inhibitory Ab (NNNIAb) ( Table 1). Altogether, these results showed that infection of pDC by HIV-1 is inhibited by NAb, but not by NNIAb. These results suggest that FccRIIa is not involved in the inhibition of pDC infection by HIV-1.
FccRIIa Does Not Participate in HIV-1 Inhibition by Antibody on Primary pDC. In order to determine the involvement of FccRIIa in Ab-mediated inhibition, the inhibitory effect NAb b12 and its mutant LALA that is unable to bind to FccRs 12 was compared on HIV-1 replication in pDC. We found that LALA and b12 similarly inhibited HIV-1 replication in pDC (Fig. 3A), indicating that HIV-1 inhibition was independent of the Fcc-FccR interaction on pDC. In addition, competition experiments were performed with anti-FccRII Ab on primary pDC. We checked that 10 mg/ml of anti-FccRII fully blocked further Ab binding (Fig. 3B). pDC pre-incubated with 10 mg/ml of anti-FccRII were infected with HIV-1 BaL in the presence or absence of NAb VRC01. We observed a similar neutralizing activity of VRC01 on HIV-1 replication whether pDC were pre-incubated or not with an anti-FccRII Ab (Fig. 3C).   Altogether, these data demonstrated that FccRIIa at the surface of pDC was not involved in the inhibition of HIV-1 by NAb.   between 20 and 600 ng/ml depending on the donors (Fig. 4). IFN-a production was strongly inhibited in the presence of anti-gp120 NAb b12 at concentrations over 10 mg/ml (Fig. 4A). The mutant version of the b12 Ab, LALA, similarly inhibited IFN-a production (Fig. 4A), indicating that FccRII was not involved in this inhibition of IFN-a production. Interestingly, NAb VRC01, also directed against the CD4 binding site of the gp120, did not inhibit IFN-a production at viral inhibitory concentrations (Fig. 4B). Similarly, anti-gp41 NAb 4E10 or most recent broadly NAb PGT121, 10-1074 and 10E8 inhibited HIV-1 replication in pDC but did not induced IFN-a production decrease (Fig. 4C and 4E). Polyclonal IgG44 purified from infected patients inhibit HIV-1 BaL infection of pDC without modifying the IFN-a production of the cells (Fig. 4D). Moreover, IFN-a secretion (100 ng/mL in control infected pDC) was not impaired following inhibition of transmitted/founder HIV-1 Bx11 infection by broadly NAb PGT121 and 10-1074 (258 and 253 ng/ mL, in PGT121 and 10-1074 treated pDC respectively). Kinetics experiments showed that IFN-a production by pDC started as soon as 6 h post-infection in the presence of HIV-1 and that the inhibition of IFN-a production observed with b12 and LALA at 48 h was the result of an early decrease in the cytokine production (Fig. 5). Therefore, inhibition of IFN-a production was observed following the treatment with the NAb b12 (and LALA) but not systematically observed following inhibition of pDC infection by other NAb. The complete protection of pDC from HIV-1 infection by the NAb was dissociated from the IFN-a production of these cells. Besides, we analyzed the presence of other cytokines and chemokines (MIP-1a, MIP-1b, IL-6, and TNF-a) in the supernatants of infected pDC in the presence or absence of NAb. HIV-1 induced the production of these four cytokines/chemokines and the presence of NAb did not decrease this production (Fig. 6). Remarkably, in the presence of viral inhibitory concentrations of NAb 4E10, the production of MIP-1a and MIP-1b was slightly increased (Fig. 6). This increase was not due to the presence of this NAb by its own as 100 mg/ml of 4E10 did not induce the production of MIP-1a or MIP-1b in the absence of HIV-1 (below 5 ng/ml). The increased MIP-1a/b production started as early as 6 h post-infection as observed in kinetics experiments (Fig. 7). The production of IL-6 or TNF-a by pDC was almost unaltered by viral inhibitory concentrations of NAb VRC01, b12 or 4E10. Altogether, these data showed that pDC maintained (or slightly increased) their production of proinflammatory cytokines/chemokines MIP-1a, MIP-1b, IL-6, and TNF-a in the presence of these NAb even though they inhibited HIV-1 replication.

Discussion
In this report we analyzed the capacity of NAb and of some NNIAb to inhibit HIV-1 infection of pDC. We demonstrated that NAb, but not NNIAb, inhibited primary pDC infection by HIV-1. Interestingly, these inhibitory activities were lower than those observed for MoDC infection. We previously showed that NAb inhibitory activity on MoDC was partially attributed to an Fc-mediated inhibitory activity involving the binding of Abs to FccRII on MoDC 17 . MoDC express both FccRIIa and FccRIIb 17,18 . The specific expression of FccRIIa on a subset of primary pDC allowed deciphering the role of the activator FccRIIa on HIV-1 inhibition. We demonstrated that FccRIIa was not involved in FccR-mediated inhibition of HIV-1 replication in primary pDC. First, the NAb b12 and its mutant version LALA, which is unable to bind to FccRs, inhibited the infection of pDC by HIV-1 with a similar efficacy. Second, blocking the FccRIIa at the surface of pDC during neutralization assays did not impede the activity of NAb VRC01. Third, NNIAb anti-PID 246-D and 4B3 that inhibit HIV-1 replication in MoDC via the participation of FccRII expressed on these cells, did not inhibit HIV-1 infection of pDC (Table 1). Consistent with our results, induction of the expression of the different FccRs in the TZM-bl cell line (used for standardized assessment of HIV-1 neutralization) has revealed that FccRIIb participates in the inhibition of HIV-1 by anti-gp41 Ab whereas the involvement of FccRIIa was negligible 32 . These results suggest a complex interplay leading to Fc-mediated inhibitory activity. Fcmediated inhibition of Abs was observed on Langerhans and interstitial DC expressing activator FccRIIa and inhibitor FccRIIb 20 . Whether primary mDC display similar Fc-mediated Ab inhibitory activity needs to be further assess. Moreover, it will be valuable to assess whether a balance between activator FccRIIa and inhibitor FccRIIb is necessary for such inhibitory activity and whether the lack of Fc-mediated inhibition in pDC was due to an impairment of FccRs functions in these cells.
pDC actively participate in early innate production of cytokines and chemokines. We analyzed the production of IFN-a, MIP-1a/b, IL-6, and TNF-a in the presence of anti-HIV-1 specific Ab. The production of MIP-1a/b by pDC was increased in the presence of HIV-1 BaL infection and NAb 4E10 compared to HIV-1 BaL alone. Previously, an increased production of MIP-1a/b has been observed in PBMC treated with anti-cardiolipin Ab 33 . The polyspecific properties of 4E10 may allow this Ab to bind to cardiolipins in addition to HIV-1 gp41 protein 34 , further accounting for this increased production of cytokines.
Remarkably, we found that the induction of IFN-a production by HIV-1 was preserved in the presence of some NAb. Indeed, IFN-a production was detected in the presence of viral inhibitory concentrations of NAb VRC01, 4E10, PGT121, 10-1074, 10E8 and polyclonal IgG44. These results contrast with the recent reports showing a decreased IFN-a production in the presence of viral inhibitory concentrations of VRC01 and 10-1074 using other experimental conditions, ie PBMC as source of pDC co-cultured with HIV-1-infected MT4C5 lymphoblastoid T cells and detection of type-I IFN with reporter cell line HL116 35 . However, we observed an inhibition of IFN-a production by NAb b12 (and its mutant LALA), consistent with previous reports using the same NAb 26 . Indeed, it was suggested that the inhibition of IFN-a production was the consequence of interaction between viral gp120 and cellular CD4 26,36,37 . Interestingly, in our study, NAb VRC01 that recognizes the CD4 binding site (CD4bs) of the gp120 38 and polyclonal IgG (that probably contains a large number of anti-CD4bs Ab) 39 , had almost no effect on IFN-a production by primary pDC while they inhibited the product-ive infection of the cells. As these Ab interact with the CD4bs of gp120, we may hypothesize that interaction leading to IFN-a production is not hindered by NAb VRC01 on primary pDC. Indeed, NAb VRC01 binds to a CD4bs epitope distinct to that of b12 and do not induce the shedding of gp120 like b12 40 . In addition, large amounts of noninfectious HIV-1 may contribute to IFN induction by pDC through induction of autophagy following TLR7 signaling 41 . These noninfectious viruses may be differentially inhibited by NAb depending to the epitope recognized. Inversely, Fc portion of IgG could downregulate virus-induced IFN-a production by pDC 42 . Therefore, mechanisms leading to IFN-a inhibition will depend on the nature of the stimuli induced both by cell surface receptors and by intracellular pathways 43 . IFN-a production in the presence of virus/ IgG immune complexes has been previously reported with other viruses [44][45][46] . For HIV-1 infection, additional studies will be necessary to define the specific Ab feature and the mechanism by which they will selectively inhibit or maintain IFN-a production by pDC. As IFN-a displays antiviral function, it may participate in HIV-1 inhibitory activity. In this study, the protective benefit of IFN-a released in the supernatant could not been assessed as primary pDC can be maintained only in culture for a few days. However, in vivo, proinflammatory cytokines released by pDC may contribute to HIV-1 elimination in addition to direct inhibition of HIV-1 replication by NAb. Whether IFN-a production by pDC during the chronic phase is detrimental or may contribute to pathogenesis is still a matter of debate [47][48][49][50] . Recent studies however suggest that IFN-a release during the acute phase may not play a direct pathologic role in HIV-1 disease progression 47,48 .
We showed here that the polyclonal IgG44, as well as anti-HIV-1 NAb VRC01, 4E10, PGT121, 10-1074, 10E8, but not b12, failed to decrease the IFN-a production induced by HIV-1 at viral inhibitory concentrations. These findings suggest that, in addition to direct inhibition of HIV-1 replication by NAb, the maintenance of proinflammatory cytokines and chemokines released by pDC may contribute to HIV-1 elimination by triggering the innate immune responses and thus impeding early HIV-1 dissemination. Therefore, in the context of a prophylactic vaccine development, induction of anti-HIV-1 NAb that simultaneously protect HIV-1 target cells and allow innate immune cells to maintain an antiviral immune response, including IFN-a production, should be favored.
Virus Preparation. Virus stocks for each primary isolate were obtained by amplification on PBMC from seronegative donors, as previously described 17,21 . The HIV-1 BaL (subtype B, R5 strain) was provided by Drs. S. Gartner, M. Popovic, and R. Gallo from the National Institutes of Health (NIH). This virus was only amplified on PBMC and therefore considered as a primary isolate. HIV-1 TV1 (subtype C, R5 strain) primary isolate was obtained from Dr. S. Engelbrecht (NHLS, Tygerberg, South Africa). Transmitted/Founder HIV-1 Bx11 primary isolate was obtained before seroconversion from French HIV-infected individual as previously described 31 . HIV-1 Specific Antibodies. Monoclonal anti-HIV-1 human Ab to gp120 IgG1 b12, LALA and PGT121 were kindly provided by Dr. D. Burton and Dr. P. Poignard (The Scripps Research Institute, La Jolla, CA). Monoclonal anti-HIV-1 human Ab to gp120 IgG1 10-1074 and gp41 10E8 were kindly provided by Dr. H. Mouquet (Institut Pasteur, Paris). Human monoclonal Ab 447-52D (anti-gp120), and 246-D (anti-gp41) were kindly provided by Dr. S. Zolla-Pazner (New York University School of Medicine, NY). Human monoclonal Ab 4E10, 2F5, 4B3, 3D6 (anti-gp41), and 2G12 (anti-gp120) were kindly provided by Dr. D. Katinger (Polymun Scientific GmbH, Vienna, Austria). Human monoclonal Ab VRC01 (anti-gp120) was kindly provided by Dr. J. R. Mascola (NIH, Bethesda, MD). Human polyclonal IgG were purified from sera of asymptomatic HIV-1 patients (approval obtained from the Comité Consultatif pour la Protection des Personnes dans la Recherche Biomédicale [CCPPRB], and informed consent was provided according to the declaration of Helsinki) 52 . And the methods were carried out in accordance with the approved guidelines.
HIV-1 Inhibitory Assays. Aliquots of 25 ml of HIV primary isolate were incubated with 25 ml of Ab for 1 hour (h) then 25 ml of either pDC (10 3 10 6 cells/ml) or MoDC (15 3 10 6 cells/ml) was added. As pDC do not support long-lasting in vitro cell culture, a single cycle infection protocol with high titer inoculums was chosen to assess virus replication in these cells. Therefore, HIV-1 primary isolates were used at 1.25 mg/ml and 2 mg/ml of p24 antigen with pDC and MoDC respectively, in order to detect about 3% infected cells. When indicated, HIV-1 reverse transcriptase inhibitor Azidothymidine (AZT, Sigma-Aldrich) was added at 5 mM to the cell culture to inhibit viral replication, allowing the residual p24 to be measured in the absence of de novo viral synthesis. After 48h, various cell surface markers and intracellular HIV-1 p24 antigen protein were stained to characterize the cells (pDC as CD123 1 , and MoDC as CD209 1 ) and to determine the percentage of infected cells by flow cytometry. The concentration of Ab required to inhibit 90% of cell infection (IC 90 ) was determined 17 . To analyze the role of FccRs in HIV-1 inhibition, pDC were incubated with 10 mg/ml of purified anti-FccRII Ab (BD Pharmingen) for 30 minutes prior to addition to the HIV-1/Ab immune complexes.
Immunophenotyping and Flow Cytometry Analysis. Briefly, cells were labeled with the Live/Dead Fixable Dead Cell Stain kit (Invitrogen/Applied Biosystems, CA) for 10 minutes at room temperature. Mouse monoclonal fluorescent Ab against human CD123-Pe-Cy5 (BD Pharmingen, USA), anti-human CD32-FITC (BD Pharmingen), anti-human CD209-PerCP-Cy5.5 (BD Pharmingen) and anti-human CD83-APC (Miltenyi Biotec) were added and the samples incubated for 10 minutes at 4uC. Cells were fixed and permeabilized using the Cytofix and Perm/Wash solutions (BD Biosciences). Then, intracellular p24 antigen was stained using anti-HIV-1 core protein p24 RD-1 (Beckman Coulter). Multi-color samples were acquired on a LSRII SORP cytometer (BD Biosciences) calibrated using Cytometer Setup & Tracking beads (BD Biosciences) to ensure consistency of fluorescence intensity measurement throughout all experiments. Compensation was performed with a CompBeads kit (BD Biosciences). Doublet cells were excluded using Forward Scatter Width and Forward Scatter Area and dead cells were excluded using Live/Dead staining. FACSDiva TM software version 6.1.2 (BD Biosciences) was used for the final analysis and graphical output.
Cytokines and Chemokines Analysis. Supernatants were collected at various time points during inhibitory assays for analysis of cytokines and chemokines production. IFN-a was measured using Verikine Human Interferon Alpha MultiSubtype ELISA kit (PBL InterferonSource, NJ). MIP-1a, MIP-1b, IL-6 and TNF-a were measured with Cytometric Bead Array (CBA, BD Biosciences) using appropriate Flex sets and Human Soluble Protein Master Buffer kit (BD Biosciences), acquired on a LSRII SORP cytometer, and analyzed using FCAP Array software version 1.0.1 (Soft Flow, Hungary).
Antibody Binding Assay. pDC were incubated with or without 10 mg/ml of purified anti-FccRII (CD32) Ab (BD Biosciences) for 30 minutes at 37uC, then labeled with anti-CD32 FITC (BD Biosciences) for 10 minutes at 4uC. The samples were fixed in Cytofix solution, acquired by LSRII SORP cytometer and analyzed using FACSDiva TM software version 6.1.2.
Statistical Analysis. Groups were compared by one-way ANOVA (Kruskal-Wallis test). Values of P , 0.05 were considered to be statistically significant. All statistical calculations were performed using GraphPad Prism 5.04 software (GraphPad, CA).