(a) Dual-label flow cytometric analysis of NK cell proliferation (decrease in CFSE fluorescence) and NKR expression gated on CD16/CD56⁺CD3⁻ cells. NK cells were stained after 5 days of coculture with uninfected (top) or HIV-1-infected (bottom) MDDC. A representative experiment showing the expression of three NK cell receptors differently modulated by HIV-1-infected MDDC, CD85j (down-regulated), CD161 (up-regulated) and NKp30 (not significantly changed). (b) Overall modifications in the percentage of NKR expression on NK cells after coculture with HIV-1-infected MDDC. Results are expressed as the percentage decrease (negative values) or increase (positive values) in NK cells expressing the different receptors after coculture with HIV-infected MDDC in comparison to coculture with uninfected MDDC. Means+SE of 9 independent experiments. * p<0.05, **p<0.01

(a) NKR expression on purified CD85j⁺ and CD85j⁻ NN cell subsets. Results are expressed as percentage of receptor expression in the given subset+SE. (b) MDDC were infected with HIV-1_Bal and then cultured alone (DC) or incubated with total NK cells (NK) or with the CD85j⁺ or CD85j⁻ NK cell subsets. HIV-1 replication was measured by p24 production at the indicated time points. Mean+SE of p24 levels in 8 independent experiments using cells from different donors. p24 levels in NK/MDDC cocultures were compared to those in MDDC cultures * p<0.02, **p<0.001. (b) MDDC were one-round infected with HIV-1/BaL pseudotypes and then cultured alone (DC) or incubated with total NK cells (NK) or with the CD85j⁺ or CD85j⁻ NK cell subsets. Ninety-six hours later, cells were lysed and luciferase activities were measured. Results are expressed as means±SE of luciferase activity of 4 independent experiments. CD85j⁺ NK cells induced higher levels of inhibition than CD85j⁻ NK cells. * p<0.05

(a) Cytolytic activity of total NK cells (NK), CD85j-enriched NK cells (CD85j+) and CD85j-depleted NK cells (CD85j−) derived from 5-day cocultures with uninfected MDDC (top panels) or HIV-1-infected MDDC (bottom panels) were analyzed for their lysis of HIV-1/GFP infected MDDC. Lysis of target MDDC replicating HIV (GFP+) and not replicating HIV (GFP-) are measured by caspase staining (quadrants B1 and B2, respectively). A representative experiment is shown. (b) Percentages of lysis of HIV-1/GFP+ MDDC target cells by total NK cells or by the CD85j⁺ or CD85j⁻ subset coculture with uninfected MDDC (HIV-, left panel) or with HIV-1-infected MDDC (HIV+, right panel). Results of 4 independent experiments are summarized. Boxes and whisker plots show the percentiles and median distribution of caspase-stained MDDC. Significant differences between NK cells from cocultures with uninfected MDDC and NK cells from cocultures with HIV-1 _Bal-infected MDDC are indicated, **p<0.01.

(a) NK cells (NK) or the CD85j⁺ or CD85j⁻ NK cell subsets were added to HIV-1/GFP-infected MDDC 2 hours after infection. (b, c) NK cells or the CD85j⁺ or CD85j⁻ NK cell subsets derived from 5-day cocultures with HIV-1_Bal-infected MDDC (b) or uninfected MDDC (c) were added to HIV-1/GFP-infected MDDC. Kinetics of HIV-1/GFP replication are expressed as percentages of MDDC cells expressing GFP (means+SE) at the indicated time points in eight independent experiments (a) or three independent experiments (b,c) Significant differences between HIV-1/GFP-infected MDDC cultures without NK cells (DC) and MDDC/NK cocultures on day 14 are indicated * p<0.05, **p<0.001.

Total NK cells or the CD85j⁺ or CD85j⁻ subsets were cocultured with HIV-1_Bal-infected MDDC in the bottom of Transwell wells. HIV/GFP-infected MDDC were added to the inserts and GFP expression was monitored over time. The percentages of NK cells expressing GFP in the inserts placed on the different cell cultures are shown: MDDC cultures without NK cells (DC), with total NK cells (NK), with CD85j⁺ NK cells (CD85j⁺) and with CD85j⁻ NK cells (CD85j⁻). Means+SE of 4 independent experiments. Significant differences between HIV-1/GFP-infected MDDC in transwell on MDDC/NK cocultures and on MDDC alone * p<0.05.

(a) MDDC were infected with HIV_Bal and incubated with CD85j⁺ NK cells after blocking CD85j receptor interactions with HLA class I molecules on the MDDC by treatment with Mabs to HLA class I ABC (HLA-ABC), E (HLA-E) or G (HLA-G) or by incubating MDDC with a recombinant CD85j protein (rCD85j). Blocking reagents were also added during coculture. MDDC infection was evaluated by measuring p24 in the supernatants 10 days later. The results are expressed as the percentage suppression of p24 (mean+SE) in comparison with HIV_Bal-infected MDDC not exposed to NK cells in 8 independent experiments. Significant differences in the levels of p24 suppression after incubation with the three anti-HLA class I Abs, with rCD85j or with rCD8j more the three anti-HLA class I Abs in comparison to CD85j⁺ NK cells cultured with infected DC without Abs are indicated by asterisks: * p<0.05, **p<0.001 (b) Flow cytometric analysis of rCD85j binding to uninfected MDDC (top panels) or to HIV_Bal-infected MDDC (bottom panels). MDDC staining with rCD85j as revealed by anti-hIgG-PE Abs (A and D). MDDC staining with a mix of anti-HLA Class I-PE mAbs after incubation with rCD85j (B and E), or with rCD85j and anti-hIgG-PE Abs after incubation with a mix of anti-HLA Class I MAbs. A representative experiment of three is shown. Without rCD85j blockade, nearly all MDDC were stained by anti-HLA-PE (not shown). The concentrations of anti-HLA class I Abs used in blocking experiments (10 µg/ml of each anti-HLA class I A,B,C, E and G) totally prevented the detection of HLA class I molecules on MDDC by using labelled anti-HLA class I Abs (not shown). Percentages of positive cells and MFI (italic) are shown. Range of positive rCD85j cells was between 8.2 and 21.9.