Downregulation and/or release of NKG2D ligands as immune evasion strategy of human neuroblastoma

Abstract

Neuroblastoma (NB) is a pediatric extracranial tumor characterized by downregulation of human leukocyte antigen class I and defects of the antigen processing machinery, two features that make it an appropriate target for natural killer (NK)-mediated lysis. NKG2D is an activating immunoreceptor expressed by cytotoxic T lymphocytes and NK cells. The ligands for NKG2D are the major histocompatibility complex class I-related chain (MIC)A and MICB glycoproteins, and the UL-16-binding proteins (ULBPs). Here, the expression of NKG2D ligands was investigated in human primary NB tumors and cell lines because scanty information is available on this issue. MICA, MICB, and ULBP transcripts were found in most tumors and cell lines. MICA protein was detected in some NB cell lines but not in primary tumors. A soluble form of MICA (sMICA) was identified in most patient sera and in some cell line supernatants. sMICA downregulated surface NKG2D in normal peripheral blood CD8(+) cells and decreased NK-mediated killing of MICA(+) NB cells. MICB was detected exclusively in the cytosol of primary tumors and cell lines. Approximately 50% of primary tumors expressed ULBP-2, but not ULBP-1 or -3. ULBP-3 was expressed in 5 of 9 cell lines, ULBP-2 in 2 of 9, whereas ULBP-1 was never detected. These studies delineate novel potential pathways of tumor escape and immunodeficiency in NB.

Figure 1

MICA, MICB, ULBP-1, -2, and -3 gene expression in primary GD2⁺ NB cells as assessed by RT-PCR. In panel A, the results obtained with two representative tumors out of the eight analyzed are shown. Neuroblasts were isolated as GD2⁺ cells by immunomagnetic bead manipulation. From left to right: MW = molecular weight markers; NC = negative control, represented by water in place of cDNA; PC = positive control, represented by 1) the GD2^- cell fraction isolated from primary NB tumors for CD45 and DRβ gene expression, 2) HeLa cells for MICA gene expression, 3) U937 cells for MICB gene expression, and 4) 293T cells for ULBP gene expression; Pt 1 = NB cells from patient 1; Pt 2 = NB cells from patient 2. The first upper panel shows the amplification product of the G3PDH housekeeping gene tested as control. On the right side of each panel, the expected MWs of the amplified bands are shown. Panel B summarizes the results obtained from the study of the eight primary NB tumors.

Figure 2

MICB and ULBP-2 protein expression in primary human NB tumors. Immunoperoxidase staining of formalin-fixed, paraffin-embedded NB primary tumors. Panel A shows one representative staining for anti-MICB mAb (S-JJ5). Inset panel shows a representative staining of ACN NB cell line for anti-MICB mAb by immunocytochemistry. Panel B summarizes the data obtained from immunohistochemical analysis of 17 disseminated and 5 localized NB primary tumors. Panel C shows one representative staining for anti-ULBP-2 mAb (M311). Panel D shows negative control stained with an isotype- and subclass-matched irrelevant antibody. Arrows indicate neuroblasts stained by the above mAbs, arrowheads indicate negatively stained lymphoid infiltrates. Original magnification, x40.

Figure 3

Surface and intracellular expression of MICB in human NB cell lines. (Panel A) The cell lines were stained with the anti-MICB mAb (S-JJ5) and subsequently analyzed by flow cytometry as detailed in the Materials and Methods section. One representative experiment out of the three performed with similar results is shown. Gray histograms represent surface staining and black histograms represent intracellular staining for MICB. U937 cells are shown as positive control for MICB surface expression. Results are expressed as MRFI. The inset in the right upper side of the figure shows a representative histogram obtained by intracellular staining of SK-N-BE-2c cells. The gray profile indicates staining with isotype-matched control mAb, whereas the black profile refers to staining with anti-MICB mAb. Panel B summarizes the results of surface staining of the whole panel of NB cell lines with mAbs to all NKG2D ligands. (Panel C) Western blot analysis of lysates from ACN NB cells (lane 1) and U937 human leukemia cells (lane 2) tested as positive control with anti-MICB mAb (S-JJ5). The position of molecular weight markers is on the right side. One representative experiment out of the three performed is shown.

Figure 4

Levels of sMICA (pg/ml) in sera of NB patients and healthy donors (panel A). Serum samples of 20 NB patients and 51 healthy donors were tested by ELISA. Panel B shows the lowest and highest values detected in both groups together with median, mean, and standard error.

Figure 5

Modulation of NKG2D expression on peripheral blood CD8⁺ cells from normal donors by sMICA⁺ serum from a NB patient. One representative experiment out of the three performed is shown. The fluorescence intensity of NKG2D was measured on PBMCs from a healthy donor that had been untreated (control: upper dot plot panel), incubated with sMICA^- normal serum (middle dot plot panel) or with sMICA⁺ serum of an NB patient (lower dot plot panel). NKG2D expression was analyzed gating on CD8⁺ cells (boxed areas in plots on the left side of the figure). Numbers in the top right corners of plots on the right side indicate NKG2D MFI (subtracted from background). MFI of IgG1 isotype negative control was 9 (data not shown).

Figure 6

Reduction of NKG2D-mediated killing of NB cells by soluble MICA. The surface MICA⁺ ACN NB cell line was tested as target of normal activated NK cells in a 4-hour chromium release assay. NKG2D-dependent cytolytic activity was evaluated by preincubating effector cells with medium, anti-MICA mAb, sMICA⁺ serum of a NB patient, sMICA⁺ serum of the same NB patient pretreated with anti-MICA mAb bound to immunomagnetic beads, or sMICA^- normal serum. The figure shows one representative experiment out of the three performed.