The HIV-1 gp120 CD4-Bound Conformation Is Preferentially Targeted by Antibody-Dependent Cellular Cytotoxicity-Mediating Antibodies in Sera from HIV-1-Infected Individuals

ABSTRACT Recent studies have linked antibody Fc-mediated effector functions with protection or control of human immunodeficiency type 1 (HIV-1) and simian immunodeficiency (SIV) infections. Interestingly, the presence of antibodies with potent antibody-dependent cellular cytotoxicity (ADCC) activity in the Thai RV144 vaccine trial was suggested to correlate with decreased HIV-1 acquisition risk. These antibodies recently were found to recognize HIV envelope (Env) epitopes exposed upon Env-CD4 interaction. CD4 downregulation by Nef and Vpu, as well as Vpu-mediated BST-2 antagonism, were reported to modulate exposure of those CD4-induced HIV-1 Env epitopes and were proposed to play a role in reducing the susceptibility of infected cells to ADCC mediated by this class of antibodies. Here, we report the high prevalence of antibodies recognizing CD4-induced HIV-1 Env epitopes in sera from HIV-1-infected individuals, which correlated with their ability to mediate ADCC responses against HIV-1-infected cells, exposing these Env epitopes at the cell surface. Furthermore, our results indicate that Env variable regions V1, V2, V3, and V5 do not represent a major determinant for ADCC responses mediated by sera from HIV-1-infected individuals. Altogether, these findings suggest that HIV-1 tightly controls the exposure of certain Env epitopes at the surface of infected cells in order to prevent elimination by Fc-effector functions. IMPORTANCE Here, we identified a particular conformation of HIV-1 Env that is specifically targeted by ADCC-mediating antibodies present in sera from HIV-1-infected individuals. This observation suggests that HIV-1 developed sophisticated mechanisms to minimize the exposure of these epitopes at the surface of infected cells.

T he IgG class of antibodies (Abs) can mediate cellular cytotoxic effector functions, such as Ab-dependent cell-mediated cytotoxicity (ADCC), viral inhibition (ADCVI), or phagocytosis (ADCP). These immune responses are driven by the engagement of the Ab Fc region with a family of proteins, known as Fc␥ receptors (Fc␥R), at the surface of effector immune cells (1). In the case of ADCC, cross-linking of Fc␥RIII (CD16) leads to the activation of the associated ITAM-containing subunits CD3 and/or FcεRI␥, which promotes the effector cells (e.g., NK cells, macrophages, or neutrophils) to perform a cytotoxic attack on the target cell (2,3). Interestingly, there is increasing evidence that ADCC plays a role in protecting against or controlling different viral infections (4)(5)(6). Accordingly, Fc-mediated effector functions were reported to correlate with decreased viral loads or rate of disease progression in both human immunodeficiency type 1 (HIV-1) and simian immunodeficiency virus (SIV) infections (7)(8)(9)(10)(11)(12)(13)(14). Additionally, it was recently suggested that ADCC could apply a significant immune pressure on HIV-1 (15), which further supports a role for this effector function in vivo. The analysis of the correlates of protection in the RV144 vaccine trial suggested that increased ADCC activity was linked with decreased HIV-1 acquisition (16). Interestingly, Abs with potent ADCC activity were isolated from some RV144 vaccinees (17). However, little is known regarding the cellular and viral determinants that govern the susceptibility of HIV-1-infected cells to ADCC killing by Abs normally elicited during the course of HIV-1 infection.
We recently reported that HIV-1 envelope (Env) interaction with the CD4 receptor at the surface of infected cells was critical for efficient ADCC activity mediated by monoclonal Abs (MAbs) targeting CD4-induced (CD4i) Env epitopes (18). Our initial findings were recently corroborated by others (19). Importantly, we reported that multiple MAbs with potent ADCC activity isolated from RV144 vaccinees also recognized Env CD4i epitopes in a manner similar to that of the well-characterized inner domain recognizing A32 Ab (18,20). Studies by other groups suggested that Abs recognizing variable regions 1 and 2 (V1V2) elicited in some RV144 vaccinees could mediate Fc effector functions (21,22). However, little is known as to the relative contribution of inner domain CD4i to the overall Fc-mediated effector functions present in sera from HIV-1-infected individuals.

MATERIALS AND METHODS
Cells. 293T human embryonic kidney and HOS cell lines (obtained from the ATCC and NIH AIDS Research and Reference Reagent Program, respectively) were grown at 37°C and 5% CO 2 in Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum (Sigma) and 100 g/ml of penicillin-streptomycin (Wysent). CEM.NKr cells (obtained from David Evans, Harvard Medical School) were grown at 37°C and 5% CO 2 in RPMI 1640 (Invitrogen) containing 10% fetal bovine serum and 100 g/ml primocine (Invitrogen).
Ficoll density gradient isolated and cryopreserved human peripheral blood mononuclear cells (PBMCs) from healthy donors were thawed and kept at 37°C and 5% CO 2 in RPMI 1640 containing 10% FBS and 100 g/ml penicillin-streptomycin for at least 16 h before subsequent experiments.
Plasmids and site-directed mutagenesis. Mutations were introduced into the previously described pNL43-ADA-GFP.IRES.Nef proviral vector (23). To generate env mutants, the SalI-BamHI fragment of pNL43-ADA-GFP.IRES.Nef was subcloned in a pUC19 intermediate before being subjected to site-directed mutagenesis using the QuikChange II XL protocol (Stratagene). The mutated insert was then cloned back into pNL43-ADA-GFP.IRES.Nef. Mutations in nef were introduced by a two-step PCR strategy using primers having 18-nucleotide overlaps and cloned back into the proviral construct using XhoI and NcoI restriction sites. All mutations were confirmed by Sanger DNA sequencing. The codon-optimized pcDNA3.1-HIV-1 YU2 ⌬V1V2V3V5 expression construct was made by replacing the sequence encoding residues 124 to 198 from the V1/V2 loop with a sequence encoding a GG linker and the sequence encoding residues 302 to 323 from the V3 loop with a sequence encoding a GGSGSG linker (24). ⌬V5 was made by replacing residues 460 to 465 with a GSG linker into pcDNA3.1-HIV-1 YU2 ⌬V1V2V3. The D368R mutation was introduced into pcDNA3.1-HIV-1 YU2 V1V2V3V5 by site-directed mutagenesis as described above.
Sera from HIV-infected individuals. Informed consent was obtained from all study participants (the Montreal Primary HIV Infection Cohort [25,26] and the Canadian Cohort of HIV-Infected Slow Progressors [27][28][29]), and research adhered to the ethical guidelines of CRCHUM. Sera were collected during Ficoll isolation of PBMCs and conserved at Ϫ80°C. Serum aliquots were heat inactivated for 30 min at 56°C and stored at 4°C until they were used in subsequent experiments. A random-number generator (GraphPad QuickCalcs) was used to randomly select a number of sera from each cohort.
Purification of recombinant HIV-1 gp120 glycoproteins. FreeStyle 293F cells (Invitrogen) were grown in FreeStyle 293F medium (Invitrogen) to a density of 1 ϫ 10 6 cells/ml at 37°C with 8% CO 2 with regular agitation (125 rpm). Cells were transfected with a pCDNA3.1 plasmid encoding codon-optimized His 6 -tagged wild-type (wt) or mutant HIV-1 YU2 gp120 using the 293Fectin reagent as directed by the manufacturer (Invitrogen). One week later, cells were pelleted and discarded. The supernatants were filtered (0.22-m-pore-size filter) (Corning), and the gp120 glycoproteins were purified by nickel affinity columns according to the manufacturer's instructions (Invitrogen). The gp120 preparations were dialyzed against phosphate-buffered saline (PBS) and stored in aliquots at Ϫ80°C. To assess purity, recombinant proteins were loaded on SDS-PAGE gels and stained with Coomassie blue.
Cell-based ELISA. Detection of trimeric Env at the surface of HOS cells was performed by cell-based ELISA, as previously described (18,30,31). Briefly, HOS cells were seeded in 96-well plates (2 ϫ 10 4 cells per well) and transfected the next day with a cytoplasmic tail-deleted HIV-1 Env YU2 variant alone or together with a human CD4 expressor using a standard polyethylenimine transfection method. Two days later, transfected cells were washed and then incubated with 1:1,000 dilutions of sera from HIVinfected cells or 1 g/ml of relevant MAbs. Env-specific IgGs were de-tected using a horseradish peroxidase (HRP)-conjugated anti-human IgG-specific secondary Ab (Pierce) with a TriStar LB 941 luminometer (Berthold Technologies).
Viral production and infection. Vesicular stomatitis virus G (VSV-G)pseudotyped NL4.3 green fluorescent protein (GFP)-encoding ADA-Env-based viruses were produced as previously described (18). Briefly, our panel of pNL4.3-GFP-ADA-based HIV-1 proviral vectors and VSV-G-encoding plasmid were cotransfected in 293T cells by standard calcium phosphate transfection. Two days after transfection, cell supernatants were harvested, clarified by low-speed centrifugation (5 min at 1,200 rpm), and concentrated by ultracentrifugation for 1 h at 4°C at 143,260 ϫ g over a 20% sucrose cushion. Pellets were harvested in fresh RPMI, and aliquots were stored at Ϫ80°C until use. Viral preparations were normalized before infection according to reverse transcriptase activity or using a standard 50% tissue culture infectious dose (TCID 50 ) procedure using TZM-bl cells (32). Viruses were then used to infect approximately 20% to 30% of CEM.NKr cells by spin infection at 800 ϫ g for 1 h in 96-well plates at 25°C.
Flow cytometry analysis of cell surface staining, ADCC responses, and binding competition assays. For cell surface staining, infected or mock-infected CEM.NKr cells were incubated for 20 min at room temperature 48 h postinfection with 1 g/ml OKT4 (anti-CD4 Ab; 14-0048-82; eBioscience), 2 g/ml BST-2 (sc-99191; Santa Cruz), 1 g/ml 2G12 (AB002; Polymun), 1 g/ml A32 Ab (kindly given by J. Robinson), or a 1:1,000 final dilution of serum from participants in PBS. Cells were then washed once with PBS and stained with 1 g/ml goat anti-mouse and anti-rabbit (Alexa Fluor 594; Invitrogen) or anti-human (Alexa Fluor 647; Invitrogen) secondary Abs for 15 min in PBS. After one more PBS washing, cells were fixed in a 2% PBS-formaldehyde solution.
Measurement of serum-and A32-mediated ADCC was performed with a previously described assay (18,33). Briefly, CEM.NKr-infected cells were stained with viability (AquaVivid; Invitrogen) and cellular (cell proliferation dye eFluor670; eBioscience) markers and used as target cells. PBMC effectors cells, stained with another cellular marker (cell proliferation dye eFluor450; eBioscience), were then mixed at an effector/target (E/T) ratio of 10:1 in 96-well V-bottom plates (Corning). A 1:1,000 final dilution of serum or 5 g/ml of the A32 MAb was added to appropriate wells. Cocultures were centrifuged for 1 min at 300 ϫ g and incubated at 37°C for 5 to 6 h before being fixed in a 2% PBS-formaldehyde solution containing 5 ϫ 10 4 flow cytometry particles/ml (AccuCount blank particles; 5.3 m; Spherotech). Samples were analyzed on an LSRII cytometer (BD Biosciences) and acquisition was set to acquire 1,000 particles, which allows the calculation of relative cell counts. Data analysis was performed using FlowJo vX.0.7 (Tree Star). The percentage of cytotoxicity was calculated with the following formula: (relative count of GFP ϩ cells in targets plus effectors) Ϫ (relative count of GFP ϩ cells in targets plus effectors plus A32 or serum)/(relative count of GFP ϩ cells in targets) according to our previously described gating strategy (18,33).
For serum adsorption and gp120 competition assays, serum dilutions from HIV-1-infected individuals were preincubated for 30 min at room temperature with purified soluble D368R gp120 ⌬V1V2V3V5 at a concentration of 83.3 pmol/l of serum. This concentration was determined by assaying the dose-dependent reduction in gp120-coated cell staining (see Fig. S4 in the supplemental material), as previously reported (33).
CD4 and BST-2 downregulation (18). Accordingly, we recently reported that cells infected with viruses defective for both Nef and Vpu present enhanced levels of CD4 and Env at the cell surface, resulting in the exposure of Env CD4i epitopes (18). To address whether these epitopes were recognized at the surface of infected cells by sera from HIV-1-infected individuals, we infected CEM.NKr cells with a panel of NL4.3-GFP ADA-Env-based viruses encoding either wild-type (wt) or Nef-and Vpu-defective accessory proteins as previously described (18). Two days postinfection, infected cells were stained with sera from HIV-infected individuals and then fluorescently labeled (Fig. 1). Interestingly, as we previously reported for the ADCC-mediating A32 Ab (18), we observed that a threshold of both CD4 and Env must be reached at the cell surface, and Env must be able to engage with CD4 in order for Env to be detected by A32 or sera from HIV-1-infected individuals ( Fig. 1 and 2; also see Fig. S1 and S2 in the supplemental material). The majority of the sera recognized more efficiently cells infected with a virus lacking Nef and Vpu than its wild-type counterpart. Cells infected with viruses lacking Nef and Vpu (i.e., presenting high levels of CD4 and Env at the cell surface [ Fig. 2A; also see Fig. S2]) but where the ability of Env to interact with CD4 was dramatically decreased by a CD4-binding site mutation (D368R) (38,39) were poorly recognized by sera from HIV-1-infected individuals. Of note, the decreased recognition of the D368R Env variant by sera from HIV-1-infected individuals was not due to decreased levels of CD4 and/or Env at the surface of cells infected with HIV-1 lacking nef and vpu (see Fig. S2).
The impact of Env-CD4 interaction on the binding of serum IgGs from 163 HIV-infected individuals was also evaluated by cell-based ELISA, as previously described (18,30,31). Briefly, HOS cells were transfected with a cytoplasmic tail-deleted HIV-1 Env YU2 variant alone or with a human CD4 expressor as reported previously (18,30). Two days later, transfected cells were washed and then incubated with 1:1,000 dilutions of sera from HIV-infected individuals. Env-specific IgGs were detected using an antihuman IgG-specific secondary Ab. Interestingly, the recognition of the Env trimer by sera from all of the clinical categories tested was significantly increased by the coexpression of CD4, while the outer domain, recognizing Ab 2G12, was not affected (see Fig. S3 in the supplemental material), suggesting that sera from HIV-1infected individuals contain a significant portion of CD4i Abs, as previously proposed (40). The ability of sera from HIV-infected individuals to mediate ADCC requires Env-CD4 interaction. We previously described that Env-CD4 interaction modulates the susceptibility of HIV-1infected cells to ADCC by CD4i Abs (18). However, whether this was also the case for sera from a large number of HIV-1-infected individuals remains unknown. Therefore, we sought to determine if the overall ADCC activity potential in sera from HIV-infected individuals would be associated with efficient Env-CD4 interaction. Using infected CEM.NKr cells as described above, we measured serum-mediated ADCC with our previously described fluorescence-activated cell sorting (FACS)-based ADCC assay (18,33). Thus, we determined the ability of 48 randomly chosen sera (including those tested in Fig. 1B) from HIV-infected individuals to mediate ADCC against infected cells (Fig. 3). Strikingly, while cells infected with wt virus were slightly more susceptible to ADCC killing than mock-infected cells, those infected with a virus lacking Nef and Vpu were dramatically more susceptible to ADCC (Fig. 3A). Interestingly, the slight increase in ADCC killing of cells lacking Nef was similar to that obtained with cells infected with a Nef variant (L166A-L168A) unable to downregulate CD4 from the cell surface (34), further stressing the importance of CD4 downregulation to protect infected cells from ADCC. However, the increase in ADCC observed with cells lacking Nef or Vpu alone was significantly lower than ADCC levels reached with cells infected with viruses lacking both accessory proteins. Moreover, the ADCC potential of different sera correlated with their ability to recognize Env at the surface of cells infected with viruses lacking both Nef and Vpu (Fig. 3B), suggesting that the efficient recognition of HIV-1-infected cells by sera is required for their ability to mediate ADCC. Finally, introducing the CD4-binding site D368R mutation in Env dramatically decreased the sensitivity of infected cells to killing by ADCC ( Fig. 3A and C), further stressing the importance of Env-CD4 interaction for exposure of epitopes recognized by ADCC-mediating Abs normally elicited in the course of HIV-1 infection.
Antibodies targeting the V1V2V3 and V5 gp120 variable regions do not play a major role in HIV ؉ serum-mediated ADCC.
Recent reports indicate that various vaccine-elicited Abs targeting conserved (CD4i) or variable epitopes (V1V2) as well as rare broadly neutralizing Abs can mediate ADCC against HIV-1-infected cells (17,18,21,22,41). To investigate whether, aside from gp120 CD4-induced antibodies, epitopes recognized by anti-gp41, variable regions, or quaternary-dependent antibodies contributed to serum-mediated ADCC responses, we designed an antibody competition assay using purified, soluble gp120 lacking variable regions V1, V2, V3, and V5 while presenting a D368R mutation (⌬V1V2V3V5 D368R) making it unable to bind cell surface CD4 (33). Of note, variable region 4 (V4) could not be removed without affecting the structural integrity of the protein (42 and data not shown). The preincubation of sera with either full-length or ⌬V1V2V3V5 D368R gp120 recombinant proteins captured anti-Env antibodies and prevented the recognition of gp120-coated cells by serum Abs (see Fig. S4 in the supplemental material). Interestingly, this was also observed at the surface of HIV-1-infected cells where competition with the ⌬V1V2V3V5 recombinant gp120 almost completely abrogated cell surface staining of cells infected with virus unable to express Nef and Vpu (Fig.  4A) and correlated with decreased ADCC activity (Fig. 4B), indicating that a recombinant protein lacking the V1V2V3V5 variable regions is sufficient to absorb the majority of ADCC activity present in sera from HIV-1-infected individuals. fection. Fc-mediated effector functions were found to inversely correlate with viral loads or decreased disease progression in SIVinfected macaques (7-9) as well as in HIV-1-infected individuals (10)(11)(12)(13)(14). Furthermore, the analysis of immune protection corre-lates in the recent RV144 vaccine trial suggested that high levels of ADCC-mediating Abs correlated with decreased HIV-1 acquisition when combined with low plasma IgA anti-Env Ab levels (16,43), warranting further studies on the viral determinants modulating ADCC. Recent observations suggested that Vpu antagonism of BST2 was sufficient to protect HIV-infected cells from ADCC (44,45). In this study, we found no significant increase in the susceptibility of cells infected with a virus lacking Vpu to serum-mediated ADCC (Fig. 3). However, in agreement with previous work done with CD4i Abs (18), decreasing Env-CD4 interaction by introducing an Env CD4 binding site mutation (D368R) in the context of a virus lacking both Nef and Vpu was sufficient to dramatically diminish killing of infected cells by serum-mediated ADCC (Fig.  3C). This is also supported by our soluble gp120 competition assay, suggesting that gp120 CD4i Abs represent the major determinant of serum-mediated ADCC in HIV-1-infected individuals (Fig. 4), since this recombinant protein could not absorb Abs directed against gp41, quaternary-dependent Abs, or V1V2V3 and V5 variable regions. In this study, we only tested sera from HIV-1-infected individuals (not from vaccinees); therefore, we do not exclude the possibility that additional types of ADCC-mediating Abs, such as anti-V1V2, elicited through vaccination, could have therapeutic utility and/or help in preventing infection.

Renewed interest in
Furthermore, stratifying the patient sera in clinical disease progression rates (classic, rapid progressors or long-term nonprogressors, as described previously [25,26]) did not reveal any significant differences between them as to the requirement of Env-CD4 interaction to promote serum-mediated ADCC against infected cells (see Table S1 in the supplemental material).
Recent observations suggested that the angle of approach of the Ab is important in order to mediate ADCC (46). Whether anti-gp120 CD4i Abs bind Env with an angle of approach that promotes the recruitment of Fc-bearing effector cells is unclear but warrants further studies.
Altogether, these data suggest that ADCC-mediating Abs are  elicited in the course of natural HIV-1 infection and that a significant proportion of them require Env to interact with CD4 in order to be effective at mediating ADCC. This supports a major role of CD4i Env epitopes in the susceptibility of infected cells to ADCC mediated by sera from HIV-1-infected individuals. Our data suggest that HIV-1, via Nef and Vpu, tightly controls cell surface levels of CD4 and Env in order to limit the exposure of potential epitopes recognized by ADCC-mediating Abs elicited in the course of natural HIV-1 infection. Therefore, targeting the ability of Vpu and Nef to downregulate CD4 and BST-2 or strategies aimed at modifying Env conformation to expose CD4i epitopes could render HIV-1-infected cells susceptible to ADCC and have therapeutic utility.

ACKNOWLEDGMENTS
We thank Nathalie Brassard, Stéphanie Matte, and the CRCHUM Flow Cytometry Platform for technical assistance, as well as Mario Legault for cohort coordination. We are thankful for the subjects' participation and collaboration.
This work was supported by a Canada Foundation for Innovation Program Leader grant, by CIHR operating grants 119334 and 134117, by FRQS Establishment of Young Scientist grant 26702 to A.F., by CIHR catalyst grant 126630 to A.F. and M.R., and by the FRQS AIDS and Infectious Diseases Network. A.F. is the recipient of a Canada Research Chair on Retroviral Entry. M.V. was supported by CIHR doctoral research award 291485, and J.R. is the recipient of CIHR fellowship award 135349.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
We have no conflicts of interest to report.