Human immunodeficiency virus type-1 (HIV-1) evades antibody-dependent phagocytosis

Abstract

Fc gamma receptor (FcyR)-mediated antibody functions play a crucial role in preventing HIV infection. One such function, antibody-dependent phagocytosis (ADP), is thought to be involved in controlling other viral infections, but its role in HIV infection is unknown. We measured the ability of HIV-specific polyclonal and monoclonal antibodies (mAbs) to mediate the internalization of HIV-1 virions and HIV-1-decorated cells by phagocytes. To measure ADP of virions, we primarily used a green-fluorescent protein-expressing molecular clone of HIV-1JRFL, an R5, clinical isolate, in combination with polyclonal HIVIG or mAbs known to capture and/or neutralize HIV-1. THP-1 and U937 cells, as well as freshly isolated primary monocytes from healthy individuals, were used as phagocytic effector cells, and uptake of virions was measured by cytometry. We surprisingly found minimal or no ADP of virions with any of the antibodies. However, after coating virions with gp41 or with gp41-derived peptides, gp41- (but not gp120-) specific mAbs efficiently mediated phagocytosis. We estimated that a minimum of a few hundred gp41 molecules were needed for successful phagocytosis, which is similar to the number of envelope spikes on viruses that are readily phagocytosed (e.g. influenza virus). Furthermore, by employing fluorescence correlation spectroscopy, a well-established technique to measure particle sizes and aggregation phenomena, we found a clear association between virus aggregation and ADP. In contrast to virions themselves, virion-decorated cells were targets for ADP or trogocytosis in the presence of HIV-specific antibodies. Our findings indicate that ADP of virions may not play a role in preventing HIV infection, likely due to the paucity of trimers and the consequent inability of virion-bound antibody to cross-link FcyRs on phagocytes. However, ADP or trogocytosis could play a role in clearing HIV-infected cells and cells on the verge of infection.

Fig 1. Env-specific antibodies do not mediate phagocytosis of HIV-1 virions.

Internalization of HIV-1_iGFP/JR-FL was measured using U937 cells (A), THP-1 cells (B), or primary monocytes (C). Uptake of HIV-1_iGFP/NL4-3 virions was measured in U937 cells (D). All mAbs were tested at a concentration of 0.05 mg/mL and the polyclonal antibody HIVIG at a concentration of 0.2 mg/mL. PGN635 is an anti-phosphatidylserine antibody included as a positive control. Data are reported as phagocytic score (% positive cells x MFI). One-way ANOVA was used in the analysis of virus uptake in the presence of an antibody compared to the no antibody control (dotted line). P-values are indicated with an asterisk: * p ≤ 0.05; ** p ≤ 0.01, and *** p ≤ 0.001. All experiments were performed in triplicate and repeated at least three times. Data are reported as means + SEM.

Fig 2. Antibodies variably capture HIV-1 virions.

HIV-specific antibodies were tested for their ability to capture infectious HIV-1_iGFP/JR-FL by determining the quantity of bound virus that infects TZM-bl cells (A). Total virus capture (without regard to infectivity) was measured by determining p24 content of captured virus by ELISA (B). Similarly, capture of infectious (C) or total (D) HIV-1_iGFP/NL4-3 was determined. RLU refers to relative light units. All antibodies were tested at a concentration of 0.002 mg/mL. Experiments were performed in triplicate and repeated three times. Data represent means + SEM.

Fig 3. Coating virus with gp41-derived peptide or gp41 results in phagocytosis by gp41-specific antibodies.

The gp41-specific mAbs 2F5 and Z13e1 mediate the internalization of MPER peptide-coated HIV-1_iGFP/JR-FL by U937 cells (A), THP-1 cells (B), and primary monocytes (C). Similarly, using U937 cells, peptide-coated HIV-1_iGFP/NL4-3 (D) and gp41-coated HIV-1_iGFP/JR-FL (E) are targets for ADP by gp41-specific mAbs. Antibodies were tested at a concentration of 0.05 mg/mL and the polyclonal antibody HIVIG at a concentration of 0.2 mg/mL. Data is represented as phagocytic score (% positive cells x MFI). One-way ANOVA was used in the analysis of virus uptake in the presence of an antibody compared to the no antibody control (dotted line). P-values are indicated with an asterisk: * p ≤ 0.05; ** p ≤ 0.01, and *** p ≤ 0.001. Experiments were performed in triplicate and repeated at least three times. Data represent means + SEM.

Fig 4. Antibodies mediate phagocytosis of aggregated virus.

To induce aggregation, HIV-1_iGFP/JR-FL virions were opsonized with HIV-specific antibodies and subsequently incubated with a goat IgG anti-human Fc-specific antibody (0.15 mg/mL) prior to being fed to U937 cells. All mAbs were tested at a concentration of 0.05 mg/mL and the polyclonal antibodies HIVIG and IVIG at a concentration of 0.2 mg/mL. Data are represented as phagocytic score (% positive cells x MFI). One-way ANOVA was used in the analysis of virus uptake in the presence of an antibody compared to the DEN3 control mAb (dotted line; which was higher than the no antibody control). P-values are indicated with an asterisk: * p ≤ 0.05; ** p ≤ 0.01, and *** p ≤ 0.001. Experiments were performed in triplicate and repeated at least five times. Data are reported as means + SEM.

Fig 5. Fluorescence correlation spectroscopy reveals quantitative differences in virion aggregation.

(A) Normalized experimental average ACFs recovered for native virus opsonized with anti-gp41 mAbs or HIVIG (A-gp41) or with anti-gp120 mAbs or HIVIG (A-gp120), gp41-coated virus opsonized with anti-gp41 mAbs or HIVIG (B-gp41) or with anti-gp120 mAbs or HIVIG (B-gp120), aggregated virus opsonized with anti-gp41 mAbs or HIVIG (C-gp41) or with anti-gp120 mAbs or HIVIG (C-gp120). (B) Normalized experimental average ACFs recovered for control groups consisting of the same virus preparations as in (A) but with no antibody or anti-dengue mAb. In both panels (A) and (B), only half of the total data points is shown for the sake of visual clarity, and the ACFs global fit to eq. S.4 (S1 Text) is shown with red continuous lines (all best-fit parameters are reported in S1 Table). (C) Hydrodynamic radius R₂ of the aggregate population recovered for each group. Radii were computed based on the Stokes-Einstein equation (S1 Text) from the best-fit estimates of the diffusion coefficient D₂. Thick error bars correspond to the standard deviations quantified by the rigorous error analysis of the ACF global fit as described in Materials and Methods. The same rigorous error analysis has also been exploited to yield, for each hydrodynamic radius, the minimum and maximum allowed values (thin error bars) compatible with one standard deviation of the ACF’s global-fit chi-square value. Results of the one-way ANOVA statistical analysis on the recovered radii are reported in S2 Table. All FCS experiments were conducted twice.

Fig 6. CEM.NKr-CCR5 cells decorated with HIV-1_iGFP/JR-FL are targets for ADP or trogocytosis by Env-specific antibodies.

HIV-1_iGFP/JR-FL virions were spinoculated onto CEM.NKr-CCR5 target cells at 4°C, which were then opsonized with HIV-specific antibodies. THP-1 effector cells (effector:target ratio = 1:1) were next incubated with the target cells for one hour. The THP-1 cells were identified by PE-conjugated anti-CD32 staining, and double positive (PE+/GFP+) cells were analyzed using flow cytometry. Antibodies were tested at a concentration of 0.05 mg/mL and the polyclonal antibody HIVIG at a concentration of 0.2 mg/mL. Data are represented as phagocytic score (% positive cells x MFI). Note that these methods do not distinguish whether target cells themselves are internalized (phagocytosis) or whether membranes containing virus are internalized (trogocytosis). One-way ANOVA was used to analyze differences compared with the no antibody control (dotted line), after demonstrating a normal data distribution. P-values are indicated with an asterisk: * p ≤ 0.05; ** p ≤ 0.01, and *** p ≤ 0.001. Experiments were performed in triplicates and repeated at least four times. Data are means + S.E.M.