Colocalization of the IL-12 receptor and FcgammaRIIIa to natural killer cell lipid rafts leads to activation of ERK and enhanced production of interferon-gamma

Abstract

Natural killer (NK) cells express an activating receptor for the Fc portion of IgG (FcgammaRIIIa) that mediates interferon (IFN)-gamma production in response to antibody (Ab)-coated targets. We have previously demonstrated that NK cells activated with interleukin-12 (IL-12) in the presence of immobilized IgG secrete 10-fold or more higher levels of IFN-gamma as compared with stimulation with either agent alone. We examined the intracellular signaling pathways responsible for this synergistic IFN-gamma production. NK cells costimulated via the FcR and the IL-12 receptor (IL-12R) exhibited enhanced levels of activated STAT4 and Syk as compared with NK cells stimulated through either receptor alone. Extracellular signal-regulated kinase (ERK) was also synergistically activated under these conditions. Studies with specific chemical inhibitors revealed that the activation of ERK was dependent on the activation of PI3-K, whose activation was dependent on Syk, and that sequential activation of these molecules was required for NK cell IFN-gamma production in response to FcR and IL-12 stimulation. Retroviral transfection of ERK1 into primary human NK cells substantially increased IFN-gamma production in response to immobilized IgG and IL-12, while transfection of human NK cells with a dominant-negative ERK1 abrogated IFN-gamma production. Confocal microscopy and cellular fractionation experiments revealed that FcgammaRIIIa and the IL-12R colocalized to areas of lipid raft microdomains in response to costimulation with IgG and IL-12. Chemical disruption of lipid rafts inhibited ERK signaling in response to costimulation and significantly inhibited IFN-gamma production. These data suggest that dual recruitment of FcgammaRIIIa and the IL-12R to lipid raft microdomains allows for enhanced activation of downstream signaling events that lead to IFN-gamma production.

Figure 1

Both FcγRIIIa and IL-12R localize to lipid raft domains on NK cells in response to FcR activation in the presence of IL-12. (A) Purified human NK cells were activated via FcγRIIIa ligation in the presence of huIL-12 for 5 minutes at 37°C (column iv). Control conditions consisted of untreated NK cells (column i), NK cells activated via FcR ligation alone (column ii), or NK cells activated with IL-12 alone (column iii). The green channel shows the location of FcγRIIIa and the red channel shows the distribution of IL-12R, while the blue channels represents p-Lck (a component of lipid raft microdomains). Single confocal sections of the cells were captured in multitrack. Each set of frames in a given column is a representative individual NK cell selected from 1 of 20 analyzed cells. Superimposed white-yellow patches signify areas of colocalization of FcγRIIIa IL-12R and p-Lck. (B) Detergent-insoluble cholesterol-enriched lipid rafts were prepared by Brij-98 extraction of purified human NK cells that were left unstimulated (left column) or activated via FcγRIIIa ligation in the presence of IL-12 for 5 minutes (right column) and separated on a sucrose density gradient. A total of 9 sequential fractions collected from the gradient were subjected to immunoblot analysis for p-Lck, FcγRIIIa, and IL-12R. Fractions 2, 3, and 4 correspond to detergent-insoluble lipid fractions; fractions 7, 8, and 9 correspond to detergent-soluble fractions. All results shown are representative of 3 independent experiments.

Figure 2

Disruption of lipid raft microdomains inhibits NK cell IFN-γ production in response to immobilized IgG and IL-12. (A) Purified human NK cells from a healthy donor were pretreated with MβCD (a cholesterol chelator) and then plated onto flat-bottom wells that were precoated overnight with 100 μg/mL human IgG (ie, immobilized Ab) in medium containing 10 ng/mL IL-12. Control conditions consisted of NK cells cultured with medium alone (Medium), immobilized IgG alone (IgG), or IL-12 alone (IL-12). NK cells were cultured with brefeldin-A for 12 hours and analyzed for IFN-γ production by intracellular flow cytometry. In another arm of the same experiment, NK cells that had been pretreated with MβCD were reconstituted with cholesterol prior to use in the immobilized IgG assay. The percentage of NK cells actively producing IFN-γ is shown within each dot plot. Results from a representative donor are shown (n = 4 donors tested). (B) NK cells pretreated with MβCD were used in an immobilized IgG assay with 10 ng/mL IL-12, as described in panel A. Cell culture supernatants were harvested after 24 hours and analyzed for IFN-γ content by ELISA. Results depict the mean plus or minus SEM of 4 donors tested. *P < .005 versus mock pretreatment (Mock Pre-Tx) control for the same stimulation condition.

Figure 3

Costimulation of NK cells leads to enhanced activation of Syk and ERK, but not p38, Jnk, or Akt. (A) Purified human NK cells (more than 97% CD56⁺/CD3⁻) underwent FcγRIIIa cross-linking by sequential treatment with an F(ab′)₂ fragment of a mouseanti–human FcγRIIIa Ab (clone 3G8) and a secondary F(ab′)₂ goat anti-mouse Ab. Recombinant human IL-12 was added at a concentration of 10 ng/mL at 37°C. Control conditions consisted of control Ab-treated NK cells (−), NK cells activated via FcR ligation alone (FcR), or NK cells activated with IL-12 alone (IL-12). Following stimulation, NK cells were lysed at the indicated time points for immunoblot analysis of p-Syk, p-ERK, p-STAT4, or p-Akt. Numbers below immunoblots represent the fold increase in phosphoprotein levels for each condition, normalized to total protein and expressed in relative densitometric units. (B-D) Time course for the activation of Syk, ERK, and STAT4 in NK cells stimulated as described in panel A. Levels of each phosphoprotein were quantified via densitometry and plotted as fold increase in phosphoprotein normalized for total protein versus time for each treatment condition. Immunoblots depict results from one representative donor. Graphs below immunoblots represent the mean plus or minus SEM from all 5 donors tested. *P < .05 versus the fold induction with IL-12 or FcR cross-linking alone at the time point.

Figure 4

IFN-γ production by FcR-stimulated NK cells in the presence of IL-12 is dependent upon activated Syk, PI-3K, and the ERK MAPK, but not p38, Jnk, or Akt. NK cells were pretreated with increasing concentrations of biochemical inhibitors of (A) Syk (piceatannol), (B) PI3-K (wortmannin), (C) ERK (U0126), or (D) Akt (Akt inhibitor; Calbiochem), and then plated in the immobilized IgG + IL-12 assay. Culture supernatants were harvested after 24 hours and analyzed for IFN-γ content by ELISA. Cells were counted via trypan blue exclusion following inhibitor pretreatment and following the IgG + IL-12 culture period to ensure equal viability of control-treated and inhibitor-treated NK cells. Pretreatment of NK cells with inhibitors of the p38 MAPK or Jnk MAPK had no effect on IFN-γ production in response to immobilized IgG and IL-12 (data not shown; similar results as with Akt inhibitor). (E) NK cells were pretreated with the indicated inhibitor and activated with various control stimuli (5 nM IL-2 or 50 ng/mL PMA plus 500 ng/mL ionomycin) for 10 minutes at 37°C. Cells were lysed, and levels of each phosphoprotein and total protein were measured by immmunoblot analysis. *P < .05 versus DMSO control for the IgG + IL-12 stimulation condition. Results represent the mean plus or minus SEM from n = 5 determinations.

Figure 5

Inhibition of PI3-K or the Syk protein tyrosine kinase within NK cells prevents ERK activation following FcR and IL-12 stimulation. NK cells were pretreated overnight with 25 μM picetannol (Syk inhibitor) or wortmannin (PI3-K inhibitor) and then activated via FcγRIIIa cross-linking in the presence of 10 ng/mL IL-12 for 5 minutes at 37°C. Whole-cell lysates were probed for levels of p-ERK and total ERK, as indicated. Numbers below immunoblots represent the fold increase in phosphoprotein levels for each condition, normalized to total ERK and expressed in relative densitometric units. All results are representative of n = 5 healthy donors.

Figure 6

Overexpression of ERK1 within primary human NK cells increases IFN-γ production in response to immobilized IgG and IL-12. Primary human NK cells were transduced with wild-type ERK1 or a DN form of ERK1 using the PINCO retroviral transfection system. Following infection, NK cells were sorted on the basis of GFP expression. (A) Representative flow plots from NK cells transduced with the empty vector or with wild-type ERK1 prior to sorting and following sorting for GFP⁺ cells. (B) Equal numbers of sorted NK cells were cultured separately on immobilized IgG in the presence of 10 ng/mL IL-12. Culture supernatants were harvested at 24 hours for quantification of IFN-γ release by ELISA. (C) Real-time PCR of sorted NK cells for ERK mRNA expression. Results represent the mean plus or minus SEM from n = 3 independent experiments. *P < .01 versus vector control cells.

Figure 7

Inhibition of lipid raft microdomains by cholesterol depletion severely attenuates activation of Syk, ERK, and STAT4 in NK cells following FcR + IL-12 stimulation. Purified human NK cells were pretreated with 5 nM MβCD (right column) or DMSO solvent control (Mock Pre-Tx; left column) for 1 hour and then activated via FcγRIIIa ligation in the presence of 10 ng/mL IL-12 for 5 minutes at 37°C. Control conditions consisted of control Ab-treated NK cells (−), NK cells activated via FcR ligation alone (FcR), or NK cells activated with IL-12 alone (IL-12). Cells were lysed and analyzed for p-Syk (A), p-ERK (B), and p-STAT4 (C) by immunoblot analysis. Numbers below immunoblots represent the fold increase in phosphoprotein levels for each condition, normalized to total protein and expressed in relative densitometric units. All results are representative of n = 3 determinations.