A novel assay for antibody-dependent cell-mediated cytotoxicity against HIV-1- or SIV-infected cells reveals incomplete overlap with antibodies measured by neutralization and binding assays

Abstract

The resistance of human immunodeficiency virus type 1 (HIV-1) to antibody-mediated immunity often prevents the detection of antibodies that neutralize primary isolates of HIV-1. However, conventional assays for antibody functions other than neutralization are suboptimal. Current methods for measuring the killing of virus-infected cells by antibody-dependent cell-mediated cytotoxicity (ADCC) are limited by the number of natural killer (NK) cells obtainable from individual donors, donor-to-donor variation, and the use of nonphysiological targets. We therefore developed an ADCC assay based on NK cell lines that express human or macaque CD16 and a CD4(+) T-cell line that expresses luciferase from a Tat-inducible promoter upon HIV-1 or simian immunodeficiency virus (SIV) infection. NK cells and virus-infected targets are mixed in the presence of serial plasma dilutions, and ADCC is measured as the dose-dependent loss of luciferase activity. Using this approach, ADCC titers were measured in plasma samples from HIV-infected human donors and SIV-infected macaques. For the same plasma samples paired with the same test viruses, this assay was approximately 2 orders of magnitude more sensitive than optimized assays for neutralizing antibodies-frequently allowing the measurement of ADCC in the absence of detectable neutralization. Although ADCC correlated with other measures of Env-specific antibodies, neutralizing and gp120 binding titers did not consistently predict ADCC activity. Hence, this assay affords a sensitive method for measuring antibodies capable of directing ADCC against HIV- or SIV-infected cells expressing native conformations of the viral envelope glycoprotein and reveals incomplete overlap of the antibodies that direct ADCC and those measured in neutralization and binding assays.

Fig 1

CD16 expression on cloned NK cell lines versus primary NK cells. (A) CD16 expression on the parental KHYG-1 NK cell line (solid gray) and the limiting-dilution clones of KHYG-1 cells transduced with retroviral vectors expressing human CD16 (red) or rhesus macaque CD16 (blue). (B) Expression levels on the human CD16⁺ KHYG-1 clone (solid red) versus levels on primary CD56^dim NK cells from 5 people (colored lines). (C) Expression levels on the rhesus CD16⁺ KHYG-1 clone (solid blue) versus levels on primary CD56^−/dim NK cells from 5 macaques (colored lines). The populations of primary NK cells shown are CD20⁻, CD3⁻, HLA-DR⁻, CD8⁺, and NKG2A⁺.

Fig 2

Luciferase activity as an indicator of cytotoxicity and target cell membrane integrity. Relative light units (RLU) indicate luciferase activity. Wells containing infected target cells and NK cells with no plasma define 100% RLU, whereas wells containing uninfected target cells and NK cells define 0% RLU. (A) Luciferase activity was lost over a 12-h time course from target cells infected with SIV_mac239 (filled symbols) in the presence of the NK cell line and plasma from SIV_mac239-infected (red) but not from naïve (blue) animals, concomitant with an increase in the low level of luciferase activity detectable in supernatants (open symbols). (B) ADCC activity against target cells infected with single-cycle SIV-eGFP was quantified by measuring luciferase activity or the frequency of cells positive for eGFP but negative for an amine-reactive dye. The dashed line indicates 50% maximal RLU or eGFP⁺ cells. (C and D) Uninfected target cells incubated with NK cells and no plasma were eGFP negative (C), whereas infected target cells incubated under the same conditions were eGFP⁺(D). (E to L) The frequencies of dead and living eGFP⁺ target cells were determined after an 8-h incubation in the presence of NK cells and plasma diluted (fold) 10² (E), 10^2.5 (F), 10³ (G), 10^3.5 (H), 10⁴ (I), 10^4.5 (J), 10⁵ (K), or 10^5.5 (L).

Fig 3

Cytotoxicity is Env-specific, CD16-dependent, and mediated by IgG. (A and B) The luciferase activity in cells infected with SIV_mac239 or SHIV_KB9 was measured after an 8-h incubation in the presence of the macaque CD16⁺ NK cell line and serial dilutions of plasma from macaques infected with SIV_mac239 (A) or SHIV_KB9 (B). The dashed line indicates 50% ADCC activity. (C and D) The parental KHYG-1 cell line was compared to the NK cell clones expressing macaque CD16 (C) or human CD16 (D) for cytotoxicity directed by SIV_mac239-infected macaque plasma against SIV_mac239-infected cells or by HIV-1⁺ patient plasma against HIV-1_NL4-3-infected cells, respectively. (E) Plasma and IgG purified from the same sample using protein A beads were compared for their ability to direct the killing of SIV_mac239-infected cells. (F) Cytotoxicity against HIV-1_NL4-3 was compared for HIV-1⁺ patient plasma versus IgG purified from the same plasma pool on protein A beads. These plasma and IgG samples were normalized for binding to SIV_mac239 or HIV-1_HxB2 gp120 in ELISAs. (G) EndoS-treated IgG from an SIV_mac239-infected macaque was compared with the untreated IgG samples for killing of target cells infected with SIV_mac239. (H) A Coomassie stain (Coo.) of total protein (left 2 lanes) and a Western blot using Lens culinaris agglutinin (LCA) to detect the Fc region N-linked glycan (right 2 lanes) were performed to confirm its removal by EndoS. Inf., infected.

Fig 4

Antibody reactivity with target cells. (A) ADCC activity against uninfected target cells was measured at a 1:32 dilution of plasma from 105 rhesus macaques. (B) The baseline level of ADCC activity at a 1:32 dilution of plasma from 386 naïve human volunteers was tested against target cells infected with HIV-1_92TH023. (C) ADCC activity titers of plasma samples against uninfected target cells before (gray) and after (black) 6 rounds of depletion were determined. The dashed line indicates 50% ADCC activity. (D) Pooled plasma from macaques immunized with single-cycle SIV produced in human cells had ADCC activity against cells infected with SIV_mac239 (black filled symbols) with potency similar to that of SHIV_SF162P3 (gray filled symbols). After 20 rounds of depletion, ADCC activity was detectable against cells infected with SIV_mac239 (black open symbols) but not SHIV_SF162P3 (gray open symbols). (E and F) Performing 20 rounds of depletion on plasma from 2 animals that lacked detectable ADCC activity against uninfected target cells did not appreciably reduce the ADCC activity against SIV_mac239-infected target cells. Inf., infected; Undep., undepleted; Dep., depleted.

Fig 5

Freezing medium causes a loss of cytotoxicity that can be prevented by calcineurin inhibitors. (A) ADCC before and after cryopreservation of the macaque CD16⁺ NK cell line under standard conditions and after cryopreservation in the presence of cyclosporine (CsA). (B to D) ADCC by macaque CD16⁺ NK cells taken through a freeze-thaw cycle (B), treated with freezing medium but not frozen (C), or treated with ionomycin (D) in the presence or absence of calcineurin inhibitor CsA or FK-506. The dashed line indicates 50% ADCC activity for a plasma sample from an SIV_mac239-infected macaque against SIV_mac239-infected target cells.

Fig 6

Reproducibility of the assay. (A) ADCC activity titers of HIVIG reagent against cells infected with HIV-1_NL4-3 were determined on 10 different days. The dashed line indicates 50% ADCC activity. (B) The CV was determined at each concentration of HIVIG for the data obtained on different days and among the triplicate samples tested on the same day. (C) The 50% ADCC titers ranged from 1.4 to 7.3 μg/ml, with a standard deviation of 2.5 μg/ml (indicated by error bar) and a CV of 65%.

Fig 7

Neutralization and ADCC of a primary HIV-1 isolate. (A and B) Plasma samples from one healthy volunteer (gray) and 10 HIV-1⁺ patients not taking antiretroviral drugs were tested for neutralization of HIV-1_YU2 in the TZM-bl assay (118, 132, 133) (A) and for ADCC against target cells infected with HIV-1_YU2 (B). Dashed lines indicate 50% neutralization (A) or ADCC activity (B).

Fig 8

ADCC versus neutralization of SIV_mac239, neutralization of SIV_mac251_TCLA, and binding to gp120. (A) Plasma samples from 37 unvaccinated rhesus macaques chronically infected with SIV_mac239 were tested for ADCC against cells infected with SIV_mac239 and against cells infected with SHIV_KB9 (gray). The dashed line indicates 50% ADCC activity. (B) ADCC activity against SIV_mac239-infected cells was quantified by calculating 50% ADCC titers and area under the curve (AUC) values for ADCC, and these measures were compared. (C and D) The same plasma samples were also tested for neutralization of SIV_mac239 (C), and the neutralization and ADCC titers were compared (D). The dashed line indicates 50% neutralization. (E and F) The titers for neutralization of SIV_mac251_TCLA in plasma samples were determined (E), and the neutralization titers against SIV_mac251_TCLA were compared with ADCC titers against SIV_mac239-infected cells (F). (G and H) Binding to SIV_mac239 gp120 was measured by ELISA (G), and the endpoint ELISA titers (H), where each titration curve crosses the dashed line, were compared with ADCC titers against SIV_mac239-infected cells. Neut., neutralization.