Broadly Neutralizing Antibody VRC01 Prevents HIV-1 Transmission from Plasmacytoid Dendritic Cells to CD4 T Lymphocytes

Abstract

Plasmacytoid dendritic cells (pDC) poorly replicate human immunodeficiency virus type 1 (HIV-1) but efficiently transfer HIV-1 to adjacent CD4 T lymphocytes. We found that coculture with T lymphocytes downregulates SAMHD1 expression, enhances HIV-1 replication, and increases pDC maturation and alpha interferon (IFN-α) secretion. HIV-1 transfer to T lymphocytes is inhibited by broadly neutralizing antibody VRC01 with efficiency similar to that of cell-free infection of T lymphocytes. Interestingly, prevention of HIV-1 transmission by VRC01 retains IFN-α secretion. These results emphasize the multiple functions of VRC01 in protection against HIV-1 acquisition.

TEXT

Novel broadly neutralizing antibodies (bNAb), such as VRC01, are able to inhibit a broad spectrum of human immunodeficiency virus type 1 (HIV-1) strains with a neutralization breadth of >90% (1–7). Their protective role has been extensively studied in various experimental infection models, including nonhuman primates (NHP) and humanized mice (8–13). Plasmacytoid dendritic cells (pDC) link innate and adaptive immunity and produce various cytokines and chemokines, especially type I interferons (IFN) (14–16). In addition, pDC present antigens and stimulate adaptive immune response as antigen-presenting cells, although with less efficiency than myeloid dendritic cells (mDC) (17–19). pDC were susceptible to HIV-1 infection (20, 21) with low levels of viral replication due to expression of various host restriction factors, such as SAMHD1 (22, 23). SAMHD1 restriction can be counteracted by the presence of Vpx, a viral protein found in HIV-2 or in simian immunodeficiency virus (SIV) from macaques (SIV_mac) (23, 24) but absent in HIV-1 (25, 26). Despite low HIV-1 replication in pDC, these cells efficiently transfer HIV-1 to adjacent CD4 T lymphocytes (27–29). HIV-1 transfer has been well described in immature monocyte-derived dendritic cells (MoDC) as a two-phase transfer with first a direct cell-to-cell passage of virus in trans followed by cis-infection with newly produced viruses (30, 31; reviewed in references 32–34). Here, we aim to analyze early phases of HIV-1 transfer from primary pDC to CD4 T lymphocytes and the effect of bNAb VRC01 on this transfer.

We used an HIV-1 transfer assay previously reported (31, 35) to mimic early mucosal HIV-1 infection and dissemination in which pDC bind and take up HIV-1 and transfer the virus in trans to CD4 T cells. Primary pDC isolated by BDCA-4 MicroBead kits (Miltenyi) from human peripheral blood mononuclear cells (PBMC) were incubated for 2 h with 500 ng/ml of primary HIV-1_BaL isolate (NIH, MD) or transmitted/founder (T/F) primary isolate HIV-1_Bx11 (obtained before seroconversion from a French HIV-infected individual [36]). After extensive washing, autologous phytohemagglutinin (PHA; 2 μg/ml)-interleukin 2 (IL-2; 0.1 μg/ml)-activated CD4 T cells, purified by positive selection after pDC purification, and anti-HIV-1 bNAb VRC01 (kindly provided by J. R. Mascola, NIH) were added to HIV-1-loaded pDC. After 72 h, we determined HIV-1 replication in the different cell types by flow cytometry (Fig. 1A). We found HIV-1_BaL replication occurred in CD4 T cells (3.6% of CD3⁺ T cells were p24⁺), demonstrating HIV-1 transfer from pDC to CD4 T cells (Fig. 1A). These percentages of p24⁺ cells correspond to newly synthesized virions, as addition of the reverse transcriptase inhibitor zidovudine (AZT) (5 μM; Sigma-Aldrich) completely abrogated the detection of p24⁺ cells (Fig. 1A). Interestingly, the percentage of infected pDC was significantly higher in the presence of CD4 T cells (8% of CD123⁺ pDC were p24⁺) than that in the absence of CD4 T cells (3% of CD123⁺ pDC were p24⁺) (Fig. 1A). An association between the percentage of HIV-1 replication in CD4 T cells and in pDC (Fig. 1B) was observed, suggesting a high degree of cooperation between CD4 T cells and pDC to promote HIV-1 replication.

FIG 1.

Measurement of HIV-1 infection and SAMHD1 expression in pDC cocultivated with autologous activated CD4 T lymphocytes. (A) The gating strategy for detection of HIV-1 replication in pDC. Among all events, forward width and forward area were used to exclude doublet cells; forward angle and side scatter light gating were used to exclude cell debris. Ab directed against human CD123 (pDC-specific surface marker) was used to select CD123⁺ pDC; Ab directed against human CD3 was used to select CD3⁺ CD4 T cells. Dead cells were then excluded with the Live/Dead fixable dead cell stain fluorescence kits (Invitrogen, CA). Percentages of living CD123⁺ pDC and CD3⁺ CD4 T cells that are infected by primary clinical HIV-1_BaL can be determined (31). Dot plots represent CD123⁺ pDC (in pink), infected with a primary HIV-1_BaL isolate or uninfected, and CD3⁺ CD4 T cells (in green) in the coculture. The HIV-1 reverse transcriptase inhibitor AZT (5 μM) was added to the coculture at the same time as CD4 T cells, as a control for HIV-1 replication. Experiments were performed in duplicate, and the mean percentages of intracellular p24⁺ pDC or CD4 T cells are shown. Productive infection was quantified by flow cytometry, based on the detection of intracellular viral p24 antigen in both cell populations after 72 h of culture. Multicolor samples were acquired on an LSRII SORP cytometer (BD Biosciences). The final analysis was performed with fluorescence-activated cell sorting (FACS) Diva software, which generated a graphical output. (B) Curve for the correlation between the mean values of percentages of infected pDC and infected CD4 T cells in coculture conditions. Pearson's correlation coefficient and its significance are shown. n = 9 experiments performed with cells from 9 healthy blood donors for panels A and B. Percentage of HIV-1-infected CD123⁺ pDC (C) and median fluorescence intensity (MFI) for SAMHD1 expression (D) in CD123⁺ pDC cocultivated with PHA–IL-2-activated CD4 T cells were measured at 72 h postinfection in the absence or presence of virus, autologous CD4 T cells, or VLP-Vpx (kindly provided by O. Schwartz, Institut Pasteur). For staining of SAMHD1 expression, anti-SAMHD1 Ab (clone I19-18) (kindly provided by O. Schwartz [35, 63]) was used following incubation with goat F(ab′)₂ fragment anti-mouse IgG1-phycoerythrin (PE) (Beckman-Coulter). The data are the means ± standard errors of the means (SEM) of pDC from 9 healthy blood donors (C) and 3 donors (D). Statistical analysis was performed using the two-tailed paired t test, and P values of <0.05 are considered significant (GraphPad Prism 5.04 software, GraphPad, California).

The increased HIV-1 replication in pDC following cocultivation with activated CD4 T cells was confirmed with n = 9 donors (P < 0.0001) (Fig. 1C). This increase was similar to that observed in the presence of virus-like particles containing Vpx (VLP-Vpx) (Fig. 1C). In parallel, we found a downregulated SAMHD1 expression in pDC (P = 0.0463) cocultivated with CD4 T cells (Fig. 1D). A decreased SAMHD1 expression was also observed in the presence of VLP-Vpx, although this difference was not statistically significant (P = 0.0899). Interestingly, SAMHD1 levels were also decreased in pDC cocultivated with autologous CD4 T cells in the absence of HIV-1 infection. These results suggest that SAMHD1 not only plays a critical role in HIV-1 restriction but may also modulate biological functions occurring during pDC-lymphocyte cross talk, as shown for MoDC (35).

As the selection of T/F virus occurs at the mucosal portal of HIV entry and disseminates through the body (37–41), the evaluation of inhibitory activity of anti-HIV-1 bNAb capable of preventing mucosal T/F HIV transmission is critical for the development of effective prophylactic and therapeutic vaccines. We thus investigated the inhibitory activity of bNAb VRC01 on the transfer of HIV-1_BaL and T/F HIV-1_Bx11 primary isolates (Fig. 2A to C and D to F, respectively). First, the relative contribution of HIV-1 transfer in trans was investigated by the addition of the protease inhibitor, indinavir (IDV; 1 μM; NIAID, NIH) (42). IDV blocks final assembly and maturation of newly synthesized virions and is therefore restricting to HIV-1 transfer in trans. In the presence of IDV, we observed a decrease of 76% and 49% of the percentages of HIV-1_BaL- and T/F HIV-1_Bx11-infected CD4 T cells, respectively, compared with control cells in the absence of IDV, indicating that 24% and 51% of infected CD4 T cells correspond to trans-infection at 72 h postinfection (Fig. 2A and D); the additional infected CD4 T cells were the results from de novo infection in cis. Moreover, 73% and 90% of pDC were infected with HIV-1_BaL and T/F HIV-1_Bx11, respectively, in the presence of IDV, compared to control-infected pDC (in the absence of IDV). Therefore, although primary pDC are less susceptible to HIV-1_Bx11 infection (about 3% of CD123⁺ pDC were p24⁺) than to HIV-1_BaL (about 8% of CD123⁺ pDC were p24⁺), a single cycle of infection was detected for both viruses at 72 h (Fig. 2A and D), allowing the investigation of the inhibition of HIV-1 transfer in trans from pDC to CD4 T cells by bNAb.

FIG 2.

Inhibition of HIV-1 transfer by bNAb VRC01. (A) Single cycle of HIV-1_BaL infection. Percentages of infection in each cocultured cell population in the presence of the HIV-1 protease inhibitor indinavir (IDV; 1 μM; NIAID, NIH) compared to control cells (in the absence of IDV). Data are expressed as the means ± SEM from n = 7 healthy blood donors. (B) The percentage of infection of pDC by HIV-1_BaL in the presence of various concentrations of VRC01 (kindly provided by J. R. Mascola, NIH) compared to control cells without VRC01 is represented. (C) IDV (1 μM) was added at the same time as VRC01 and CD4 T cells to limit HIV-1_BaL transfer to trans. After 72 h, the percentages of infected CD4 T cells (in green) and infected pDC (in pink) were determined by flow cytometry. Data are means ± SEM from at least 6 independent experiments with cells from 6 healthy blood donors. The capacity of VRC01 to inhibit HIV-1_BaL cell-free infection of CD4 T cells (with a conventional neutralization assay using autologous activated CD4 T cells, n > 4 healthy blood donors) was analyzed (in black). HIV-1 infection in primary pDC and CD4 T cells (D) and inhibitory activity of bNAb VRC01 (E and F) were assessed under the same conditions as described for panels A, B, and C in the presence of transmitted/founder HIV-1_Bx11 virus. Means ± SEM from at least 3 independent experiments performed with pDC from 3 healthy blood donors are shown. Statistical analysis was performed using the two-tailed paired t test, and P values of <0.05 are considered significant.

When the bNAb VRC01 was added at the concentration of 20 μg/ml to HIV-1-loaded pDC at the same time as activated CD4 T cells, transfer of HIV-1_BaL and T/F HIV-1_Bx11 to CD4 T cells was prevented by 87% and 78%, respectively (Fig. 2B and E). In contrast, HIV-1 replication in pDC only slightly decreased (about 20%) when VRC01 was added 2 h after incubation of pDC with HIV-1_BaL or T/F HIV-1_Bx11 (Fig. 2B and E). We further analyzed the effect of VRC01 on HIV-1 transfer in trans by adding IDV to the cocultures (Fig. 2C and F). We observed a similar inhibition of CD4 T cell infection in the presence of 20 μg/ml of VRC01, demonstrating its capacity to inhibit HIV-1 transfer in trans from pDC to CD4 T cells as well as further transfer in cis to CD4 T cells. Besides, we compared VRC01-mediated inhibition of transfer with inhibition of cell-free viral particles infecting CD4 T cells in a conventional neutralization assay. Equivalent percentages of infected T cells and similar NAb inhibitory concentrations were recorded for both assays (Fig. 2B and C or E and F). These results indicate that HIV-1 transfer from pDC to CD4 T cells is inhibited to a similar extent as cell-free virus particles in these culture conditions. A similar inhibition was previously observed between cell-free and HIV-1 transfer from MoDC to CD4 T cells (31) and confirmed in this study. Interestingly, a similar inhibitory activity of VRC01 was recently reported for the HIV-1 transmission from macrophages to T cells (43). In contrast, Abela et al. and Sagar et al. found that VRC01 did not block cell-mediated HIV-1 transmission from HIV-1-infected PBMC or mature MoDC to TZM-bl target cells and proposed that cell-to-cell transmission enabled HIV-1 to evade inhibition by anti-gp120 Ab (44, 45). Several mechanisms might be explained by the different findings observed. First, the difference of inhibitory activity may be attributed to the HIV-1 target cells used in the transfer assay. As TZM-bl overexpress CD4/CCR5 HIV-1 receptor/coreceptor entry molecules, we may hypothesize that higher concentrations of VRC01 are necessary to prevent binding to receptor/coreceptor on these cells. Second, immunological synapse formation may differ according to the type of cell used. DC-lymphocyte cross talk involves ICAM-1 and LFA-1 adhesion molecules and stabilizes interactions (33, 35, 43, 46, 47) that are absent on TZM-bl cell lines (48, 49). The strength of the synapse established and the efficiency of HIV-1 transfer may influence the inhibitory activity of HIV-1-specific NAb, making the comparison with different donor/target cell pairs difficult (50). In this regard, the close interaction between DC and autologous lymphocytes may modulate the immune response (32, 35, 51–54). Finally, the density of Env expressed on infected cells may influence the potential of NAb inhibition. In this regard, the amount of Env expressed on infected pDC may be lower than that expressed on other infected cells. Therefore, although more energy may be required to block fusion between infected pDC and uninfected T cells compared to the energy required to block fusion between free virus particles and T cells, this type of event may be limited by the amount of Env expressed on pDC, limiting the concentration of NAb required. In conclusion, the NAb activity in cell-to-cell transmission versus cell-free infection will be difficult to normalize, as it relies on the capacity to enter and infect the different cell types (reviewed in reference 55–57).

Cross talk between pDC and lymphocytes induces immune activation; the expression of the costimulation marker CD86 and DC maturation marker CD83 was therefore determined on the surface of HIV-1-infected pDC in the presence or absence of VRC01. Infection with HIV-1_BaL or HIV-1_BX11 did not induce pDC maturation, but the maturation markers were increased in the context of coculture with CD4 T cells in the presence or absence of HIV-1 and independently of VRC01 or IDV treatment (Fig. 3A, B, and D). Similar triggering of DC maturation by DC-lymphocyte cross talk was also described for MoDC (31). As pDC are the major producers of type I IFN in response to viral infection (15), IFN-α production was measured in the supernatants (Fig. 3C and E). We found that IFN-α secretion was induced following HIV-1_BaL or HIV-1_BX11 infection, and this production varied between 10 and 200 ng/ml, depending on the donors. Moreover, IFN-α production was significantly increased in the presence of CD4 T cells (Fig. 3C and E). Noteworthy, IFN-α production was not decreased following VRC01 inhibition of HIV-1 transfer or IDV treatment (Fig. 3C and E). These results contrast with the recent reports showing a decreased IFN-α production in the presence of viral inhibitory concentrations of VRC01 using other experimental conditions, i.e., PBMC as a source of pDC cocultured with HIV-1-infected MT4C5 lymphoblastoid T cells and detection of type I IFN with reporter cell line HL116 (58). Different experimental protocols, as well as different detection methods, may explain our discrepant findings. Of note, the transduction of VLP-Vpx to infected pDC failed to induce pDC maturation and IFN-α production (Fig. 3B and C), consistent with previous reports (22). Our findings indicate that pDC-lymphocyte cross talk contributes to pDC activation and innate immune sensing of HIV-1.

FIG 3.

Coculture with CD4 T lymphocytes enhances maturation and innate sensing of HIV-1-infected pDC. The percentages of CD123⁺ pDC expressing CD86⁺ (A) and CD83⁺ (B and D) were determined by flow cytometry following infection of HIV-1_BaL (A and B) or T/F HIV-1_Bx11 (D) in the presence of activated CD4 T cells and various concentrations of VRC01, IDV (1 μM), or VLP-Vpx at 72 h postinfection. (C, E) IFN-α secretion in the supernatants of pDC infected with HIV-1_BaL (C) or T/F HIV-1_Bx11 (E) was detected using the Verikine human IFN-α MultiSubtype enzyme-linked immunosorbent assay (ELISA) kit (PBL Interferon Source, New Jersey) under the same conditions as described for panel B compared to the control (HIV-1-infected pDC alone) at 72 h postinfection. At least 6 independent experiments performed with pDC from 6 healthy blood donors are shown for HIV-1_BaL, and 3 independent experiments with cells from 3 healthy blood donors are shown for T/F HIV-1_Bx11. Statistical analysis was performed using the two-tailed paired t test, and groups were compared by one-way ANOVA (Kruskal-Wallis test) with P values of <0.05 considered significant.

Recent studies showed that pDC sense hepatitis C virus (HCV)- or lymphocytic choriomeningitis virus (LCMV)-infected cells (59, 60), and viral transmission by cell-to-cell contacts are more potent inducers of IFN than infection by cell-free viral particles (61, 62). We confirmed and extended these data to HIV-1 transfer from primary pDC to CD4 T lymphocytes. Interestingly, IFN-α induced in cocultured pDC by HIV-1 and the pDC maturation was not modified following the inhibition of HIV-1 transfer by VRC01. These results suggest that increased immune sensing and DC maturation during pDC-lymphocyte cross talk could promote efficient innate immune responses and the ability to control viral infection. Moreover, we observed a significant downregulation of host restriction factor SAMHD1 in pDC following coculture with autologous CD4 T cells, which might be responsible for the increase of IFN-α production by cocultured infected pDC. Therefore, we propose that the involvement of SAMHD1 in the restriction of HIV-1 replication and in the triggering of immune responses described for MoDC (35, 63) also accounts for primary pDC.

In summary, we demonstrate that the potent and bNAb VRC01 inhibits transfer of a T/F primary isolate from primary pDC to CD4 T lymphocytes with an efficiency similar to that of cell-free infection of CD4 T lymphocytes. These findings are very encouraging observations for the in vivo role of bNAb in protection from early HIV-1 transmission and rapid dissemination in the body. Together with previous studies (31, 64, 65; reviewed in reference 55), these results suggest that at the mucosal portal of HIV-1 entry, the enhanced virus replication in pDC promotes HIV-1 transfer to neighboring CD4 T lymphocytes. HIV-1-specific NAb may prevent this early virus dissemination if locally present early after sexual transmission.