Oncogene MYCN regulates localization of NKT cells to the site of disease in neuroblastoma

Abstract

Valpha24-invariant natural killer T (NKT) cells are potentially important for antitumor immunity. We and others have previously demonstrated positive associations between NKT cell presence in primary tumors and long-term survival in distinct human cancers. However, the mechanism by which aggressive tumors avoid infiltration with NKT and other T cells remains poorly understood. Here, we report that the v-myc myelocytomatosis viral related oncogene, neuroblastoma derived (MYCN), the hallmark of aggressive neuroblastoma, repressed expression of monocyte chemoattractant protein-1/CC chemokine ligand 2 (MCP-1/CCL2), a chemokine required for NKT cell chemoattraction. MYCN knockdown in MYCN-amplified neuroblastoma cell lines restored CCL2 production and NKT cell chemoattraction. Unlike other oncogenes, MYCN repressed chemokine expression in a STAT3-independent manner, requiring an E-box element in the CCL2 promoter to mediate transcriptional repression. MYCN overexpression in neuroblastoma xenografts in NOD/SCID mice severely inhibited their ability to attract human NKT cells, T cells, and monocytes. Patients with MYCN-amplified neuroblastoma metastatic to bone marrow had 4-fold fewer NKT cells in their bone marrow than did their nonamplified counterparts, indicating that the MYCN-mediated immune escape mechanism, which we believe to be novel, is operative in metastatic cancer and should be considered in tumor immunobiology and for the development of new therapeutic strategies.

Figure 1. MYCN represses CCL2 expression and inhibits the chemoattraction of NKT and T cells.

CHLA-255 neuroblastoma cells were stably transduced with MYCN cDNA or empty vector control. (A) Representative Western blot analysis of MYCN protein in the nuclear extracts from indicated cells. (B) CCL2 RNA expression was quantified by TaqMan RT-PCR; values are relative to expression in parental cells. Results are mean ± SD from 3 experiments. (C) CBA analysis. Five indicated chemokines were measured in supernatants collected after 24 hours’ culture of 80% confluent cells. Chemokine concentration is proportional to the fluorescence intensity on the x axis and was calculated by standard curve analysis. Concentrations above 20 pg/ml are shown. (D) The same supernatants as in C were placed in lower chambers of dual-chamber plates with 5-μm pore membrane. Freshly isolated human monocyte-depleted PBLs were added in upper chambers and allowed to migrate for 4 hours. Absolute numbers of NKT and T cells were determined by flow cytometry using CD3/6B11 staining and TruCount tubes with a known number of beads (see Methods). Percent migration reflects the proportion of migrated cells relative to input cells. Specific migration was calculated by subtracting the number of cells that migrated to control medium. Results are mean ± SD from 4 experiments.

Figure 2. MYCNkd induces CCL2 expression in neuroblastoma cells and NKT cell chemoattraction.

SK-N-BE(2) and LA-N-1 MYCN-amplified neuroblastoma cell lines were stably transduced with MYCN shRNA using a U6 expression cassette into pHIV-7-EGFP lentiviral plasmid. (A) Representative Western blot analysis of MYCN protein in the nuclear extracts from indicated cells. Lane 1, parental; lane 2, vector control; lane 3, MYCN shRNA. (B) CCL2 RNA expression was quantified by TaqMan RT-PCR; values are relative to expression in parental cells. Results are mean ± SD from 3 experiments. (C) CBA analysis was performed with supernatants of vector control or MYCN shRNA-transduced SK-N-BE(2) cells as described in Figure 1C. (D) Migration of freshly isolated PBLs to the neuroblastoma cell supernatants with the indicated conditions was performed as described in Figure 1C. Where indicated, supernatants were pretreated for 1 hour with neutralizing anti-CCL2, anti-CXCL8, or isotype control (IgG1) mAb. Results are mean ± SD from 3 experiments.

Figure 3. MYCN neither activates STAT3 nor affects DC phenotype.

(A) Lysates from parental, vector-, or MYCN-transduced CHLA-255 cells were examined for phosphorylated STAT3 (p-STAT3) and for STAT3 by Western blot. Treatment with recombinant human IL-6 (10 ng/ml for 15 minutes) was used as a positive control. (B) Flow cytometric analysis of viable DCs generated in one-half volume conditioned medium from indicated neuroblastoma cells. One-half volume Iscove’s modified Dulbecco’s medium and conditioned medium from IL-6–treated (24 hours) CHLA-255 cells were used as negative and positive controls, respectively. Numbers in plots denote percent CD80⁺CD86⁺ cells. Shown are representative plots from 3 experiments.

Figure 4. MYCN directly binds CCL2 promoter.

(A) CCL2 promoter cDNA constructs with (2.8 kb) and without (2.6 kb) E-box site were fused with a Firefly luciferase cDNA. A pGL3 plasmid containing only Firefly luciferase cDNA was used as a negative control. CHLA-255 or CHLA-255/MYCN cells were transiently cotransfected with the indicated constructs and pRLSV40 plasmid containing Renilla luciferase cDNA as an internal control. The activity of CCL2 promoter was detected in a dual-luciferase assay and expressed as a ratio of Firefly to Renilla luciferase luminescence intensity. (B) Nuclear extract from neuroblastoma cells was analyzed by EMSA with biotinylated E-box probe from CCL2 promoter. Lane 1, free probe; lane 2, nonbiotin (cold) probe plus probe plus CHLA-255/MYCN; lane 3, probe plus CHLA-255/vector; lane 4, probe plus CHLA-255/MYCN; lane 5, probe plus CHLA-255/MYCN plus anti-MYCN mAb; lane 6, probe plus CHLA-255/MYCN plus IgG control. Data are representative of 6 experiments. (C) Nuclear extract from CHLA-255/MYCN was analyzed by EMSA with E-box probe as above (positive control) or with INR probes. Lane 1, free E-box probe; lane 2, E-box probe plus CHLA-255/MYCN; lanes 3, 5, and 7, free probes for INR-1, INR-2, and INR-3, respectively; lanes 4, 6, and 8, CHLA-255/MYCN plus probes for INR-1, INR-2, and INR-3, respectively. Data are representative of 3 experiments.

Figure 5. MYCN regulates NKT cell localization to neuroblastoma xenografts in BM of NOD/SCID mice.

(A) CHLA-255 and CHLA-255/MYCN neuroblastoma cells were injected into the bone cavities of left and right femurs in NOD/SCID mice, and 4 weeks later in vitro–expanded human NKT cells were injected i.v. together with human PBMCs (1:10 ratio, 5 × 10⁷ total cells). Mice were sacrificed after 24 hours, and BM from the indicated sources was analyzed by 5-color flow cytometry. After exclusion of DAPI⁺ dead cells, the neuroblastoma (NB) cells and leukocytes (Leuk) were identified as CD56⁺CD45^– and CD45⁺ events, respectively. T cells (CD3⁺6B11^–) and NKT cells (CD3⁺6B11⁺) were analyzed after gating on CD45⁺ region. BM of tumor-free tibias served as control. Numbers within parentheses and brackets represent percentages of all viable cells (DAPI^–) and of leukocytes, respectively. Data are from a representative individual mouse. ND, not detected. (B) Scatter plot of T, NKT cell, and monocyte (Mono) frequency among all non-neuroblastoma cells in BM from individual mice as indicated in 3 different experiments. Data are mean ± SEM from 16 animals.

Figure 6. CCL2 is required for leukocyte migration to neuroblastoma xenografts in BM of NOD/SCID mice.

Xenografts of CHLA-255 neuroblastoma cells were grown in femurs of NOD/SCID mice for 4 weeks. Neutralizing antibodies against CCL2, CXCL8, or isotype control (IgG) at 20 μg per mouse were injected i.p. 24 hours before adoptive transfer of human PBMCs, with subsequent analysis of leukocyte migration as in Figure 5. Data are mean ± SD from 5 experiments. Each experiment included all groups, with 1 mouse per group and PBMCs from the same donor.

Figure 7. NKT cell number is decreased in BM of MYCN-amplified neuroblastomas with BM metastases.

BM specimens were analyzed by 5-color flow cytometry as in Figure 5. Shown are scatter plots of (A) NKT cell and (B) T cell frequency distribution in BM from MYCN-amplified (A) versus nonamplified (N) neuroblastomas with more than 10% tumor cells in BM. Tumor-free BM specimens from age-matched patients with stage 4 (amplified and nonamplified) neuroblastomas were used as control. Data are mean ± SEM from BM specimens of 25 patients.