Presentation of alpha-galactosylceramide by murine CD1d to natural killer T cells is facilitated by plasma membrane glycolipid rafts

Abstract

CD1 molecules are non-polymorphic major histocompatibility complex class I-related proteins that bind and present glycolipid antigens to T-cell antigen receptors (TCR) expressed by alphabeta T cells or natural killer-like T cells (NKT). Anti-metastatic properties of NKT cells reactive to the CD1d-binding antigen alpha-galactosylceramide (alpha-GalCer) are now being explored as a contributor to tumour cell killing. In this study, we tested the hypothesis that presentation of alpha-GalCer by murine CD1d (mCD1d) to mCD1d-restricted NKT cells was facilitated by plasma membrane glycolipid rafts. Confocal microscopy of mCD1d-transfected A20 B cells (A20mCD1d) demonstrated that mCD1d was raft-localized. This observation was confirmed by immunoblotting of raft fractions isolated on sucrose density gradients. Raft disruption by the cholesterol-binding agent nystatin, or short-chain ceramides, inhibited presentation of low concentrations of alpha-GalCer to NKT cells. Inhibition of antigen presentation was reversed by treatment of A20mCD1d cells with higher alpha-GalCer concentrations, or removal of raft-disrupting agents. These data indicate that partitioning of mCD1d into membrane rafts increases the capacity of antigen-presenting cells to present limiting quantities of glycolipid antigens, perhaps by stabilizing mCD1d/antigen structures on the plasma membrane and optimizing TCR engagement on NKT cells.

Figure 1

Plasma membrane distribution of raft and non-raft markers. Cells were incubated at 4° and stained for GM-1 (red) and: mCD1d; BCR; I-Ad; TfR; and CD45 (green). (a) Cells were fixed and analysed by confocal microscopy. Red/green colocalization is shown in yellow. (b) For each marker, images of at least 100 cells were scored for red/green colocalization. Graph shows the percentage of cells in the sample demonstrating GM-1 colocalization of the indicated surface marker.

Figure 2

Nystatin reversibly decreases mCD1d colocalization to rafts. Cells were resuspended in serum-free media and treated with 75 μg/ml nystatin for 15 min and then (a) fixed in 1·0% w/v paraformaldehyde, or (b) allowed to recover in serum-containing media for 30 min, before fixation. Cells were then stained for mCD1d (green) and GM-1 (red). Images shown are representative of at least 100 cells for each condition.

Figure 3

Detection of mCD1d in murine B-cell rafts. (a) BCL1 and (b) A20mCD1d cells were subject to detergent lysis and raft isolation. Equal volumes of gradient fractions (2 × 10⁶ cell equivalents) were resolved by SDS–PAGE before transfer to nitrocellulose and immunoblotting for mCD1d, Lyn, TfR, and β-tubulin. (c) Cells were subject to detergent lysis and raft isolation, or treated with 12·5 mm MβCD for 1 hr at 37° prior to detergent lysis. Equal volumes of gradient fractions were resolved by SDS–PAGE and immunoblotted for mCD1d. (d) Cells were treated as follows: Lanes A, untreated; lane B, 12·5 mm MβCD; lanes C, 12·5 mm MβCD then recovery; lane D, 2·5 mm MβCD + 75 μg/ml nystatin; lanes E, 75 μg/ml nystatin; or lanes F, 75 μg/ml nystatin then recovery, before detergent lysis and raft isolation. Raft fractions were resolved by SDS–PAGE and immunoblotted for mCD1d.

Figure 4

Presentation of α-GalCer to NKT cells is facilitated by membrane rafts. (a, inset) Cells were pulsed with α-GalCer at the concentrations indicated for 4 hr, before fixation in 0·15% w/v paraformaldehyde and washing in NKT culture media. Cells were then cocultured as indicated for 20 hr, before collection of supernatants and measurement of IL-2 concentration by ELISA. Data shows mean ± SD IL-2 concentration in media for triplicate samples. (a, main graph) Cells were pulsed with α-GalCer at the concentrations indicated before fixation and washing (control, closed bars), or nystatin treatment (75 μg/ml), fixation and washing (open bars). Cells were then cocultured with NKT cells as described. (b) After pulsing with α-GalCer cells were untreated, or treated with 75 μg/ml nystatin. Antigen-pulsed, non-fixed A20mCD1d cells were then allowed to recover by washing in NKT culture media, before addition of NKT cells. Data shows the mean ± SD IL-2 production for triplicate samples and is representative of at least three independent experiments.

Figure 5

Recovery from nystatin treatment in fixed cells. (a) A20mCD1d cells were pulsed with α-GalCer at the concentrations indicated before addition of 75 μg/ml nystatin and fixation. (b) Nystatin-containing medium was removed and cells were allowed to recover in serum-containing medium for 30 min, before fixation. Data show the mean ± SD IL-2 production for triplicate samples from a single experiment.

Figure 6

Effects of short-chain ceramides on mCD1d raft colocalization and antigen presentation. (a) A20mCD1d cells were incubated with C2-Cer, or DH-Cer at a 100 μg/ml final concentration, before fixation and staining for mCD1d (green) and GM-1 (red). At least 100 cells were analysed by confocal microscopy. Mid-section images are shown for four representative cells. (b) Cells were treated with C2-Cer or DH-Cer for 1 hr before detergent lysis and raft isolation. Gradient fractions were resolved by SDS–PAGE and immunoblotted for mCD1d. (c) Cells were pulsed with α-GalCer at the concentrations indicated before addition of C2-Cer or DH-Cer. Cells were then fixed, washed, and cocultured with NKT cells before measurement of IL-2 concentrations by ELISA. Data show the mean of triplicate samples ± SEM. (d) The mean of three experiments performed as in (c). Data show mean ± SEM (n = 3) IL-2 production in C2-Cer-treated cells as a percentage of that in DH-Cer-treated cells.