Medulloblastoma expresses CD1d and can be targeted for immunotherapy with NKT cells

2.1. Human specimens

PBMC or frozen tumor specimens from MB patients (mean age of 7.8, range 2–16 years old) were obtained at diagnosis at Texas Children’s Cancer Center, Baylor College of Medicine or Children’s Hospital Los Angeles, respectively, according to the insitution’s approved IRB protocols. Informed consent was obtained in accordance with institutional review board policies and procedures for research dealing with human specimens. PBMC of healthy donors (at least 17 years old) were isolated by gradient centrifugation from buffy coats purchased from Gulf Coast Regional Blood Center (Houston, TX). Medulloblastoma tissue arrays from 20 unidentifiable MB patients and normal human brain tissues were purchased from Cybrdi, Inc. Molecular subgroups of MB patients were identified using quantitative reverse transcription polymerase chain reaction (qRT-PCR) and a medulloblastoma gene signature derived from prior microarray studies [5;6] and a CHLA study (manuscript in preparation).

2.2. Cell lines and culture conditions

The medulloblastoma cell lines (Daoy, D283, and D341) and 293T cells were purchased from the ATCC. MED8A line was kindly provided by Dr. R. Gilbertson (St. Jude Children’s Research Hospital, Memphis). All medullobalstoma cell lines were maintained in DMEM (Invitrogen) with 10% FCS (HyClone) supplemented with 2 mM GlutaMAX-I, 1.5 g/L sodium bicarbonate, 0.1 mM nonessential amino acids, and 1 mM sodium pyruvate (Invitrogen). NKT cell lines were expanded from PBMCs of healthy volunteers as previously described with modifications [22]. Briefly, PBMCs were isolated from buffy coats by Ficoll-Hypaque gradient density centrifugation. NKTs were purified by anti-iNKT microbeads (Miltenyi Biotec). The negative PBMC fraction was irradiated (4000 Rad) and aliquoted. NKTs were stimulated with an aliquot of autologous PBMCs pulsed with αGalCer (100 ng/mL, Funakoshi Co.Ltd). rhIL-2 (200 U/ml, BDP, National Cancer Institute Frederick) was added on the second day and then every other day. NKTs were restimulated every two weeks with the remaining PBMC aliquotes. The phenotype and purity of NKTs were assessed using mAbs for CD3, Vα24-Jα18 (6B11), and CD4. NKT cell ligand 7WD8-5 was kindly provided by Dr. Moriya Tsuji (The Rockefeller University, New York, NY). βGlcCer was purchased from Avanti Polar Lipids. All lipids were dissolved in DMSO for in vitro assays and PBS buffer (5.6% sucrose, 0.75% L-histidine and 0.5% Tween20) for in vivo experiments.

2.3. RNA isolation and Real-time RT-PCR

Total RNA from cultured cells were extracted using TRIZol reagent (Invitrogen). cDNAs were synthesized by super reverse transcript kit (Invitrogen). Quantitative RT-PCR for CD1d was performed with TaqMan gene expression assay using the ABI7900 sequence Detection System (Applied Biosystems). The relative change in gene expression was calculated based on the δCt method using HPRT1 housekeeping gene as a control.

2.4. Immunohistochemistry

Human medulloblastoma tissue array slides and normal brain paraffin tissue sections were purchased from Cybrdi, Inc. H&E staining was done following a standard hematoxylin and eosin staining protocol. CD1d and CD3 expression was detected with immunohistochemistry using mouse anti-human CD1d mAb (clone NOR3.2, Abcam) and mouse anti-human CD3 mAb (clone PS-1, Vector Laboratories).

2.5. Flow cytometry

Surface CD1d expression on cell lines was assessed using an anti-CD1d-PE 42.1 mAb. NKT cells were identified after PBMC staining with anti-NKT TCR-PE 6B11 and anti-CD3-APC mAbs. Manufacture recommended fluorochrome- and isotype-matching mAbs were used for negative controls (BD Biosciences). The analysis was performed on a LSR-II four-laser flow cytometer (BD Biosciences) using BD FACDiva software v. 6.0 and FlowJo 7.2.5 (Tree Star, Inc).

2.6. Multiplex cytokine quantification assay

Cytokines released by NKT cells were assessed by CBAPlex beads on FACSArray bioanalyzer (BD biosciences) according to the manufacture’s manual and as previously described [15].

2.7. In vitro cytotoxicity assay

Daoy.ffLuc cells were pulsed overnight with a NKT cell ligand or DMSO control and plated in a 96-well black plate at 20,000 cells/well. NKT cells were added at different effector to target cell ratios. Anti-hCD1d 51.1 blocking mAb was kindly provided by Dr. S. Porcelli (Albert Einstein College of Medicine, Bronx, NY). After 4-hr co-culture, luciferin was added in each well and luminescence was quantified by a plate reader Infinite^® M200 (Tecan). The number of viable Daoy.ffLuc cells in each well was calculated based on standard curve generated from serial dilutions of the target cells. NKT cell cytotoxicity was calculated using the following formula: cytotoxcicity % = (cell number in control well – cell number in assay well) * 100/cell number in control well (Daoy.ffLuc cell alone).

2.8 Orthotopic xenogenic model of medulloblastoma in NOD/SCID mice

All animal experiments were conducted on a protocol approved by the Baylor College of Medicine Institutional Animal Care and Use Committee (IACUC). NOD/SCID mice were purchased from Jackson Lab. Intracranial injection were conducted as previously described with modifications [12]. Male 9 week to 12 week-old mice were anesthetized with rapid sequence inhalation isofluorane (Abbot Laboratories, England) followed by an intraperitoneal injection of 225–240 mg/kg Avertin^® solution and then maintained on isofluorane by inhalation throughout the procedure. Mice were immobilized in a Cunningham^™ Mouse/Neonatal Rat Adaptor (Stoelting, Wood Dale, IL) stereotactic apparatus fitted into an E15600 Lab Standard Stereotactic Instrument (Stoelting) and scrubbed with 1% povidone-iodine. A 10 mm skin incision was made along the midline. The tip of a 31G ½ inch needle mounted on a Hamilton syringe (Hamilton, Reno, NV) served as the reference point. A 1mm burr-hole was drilled into the skull, 1 mm anterior to and 2 mm to the right of the bregma. Daoy cells (2.5 × 10⁵ in 2.5 μL) were injected 3 mm deep to the bregma, corresponding to the center of the right caudate nucleus over 5 minutes. The needle was left in place for 3 minutes, to avoid tumor cell extrusion, and then withdrawn over 5 minutes. Five days after tumor cell injection, animals received 2 × 10⁶ NKT lymphocytes alone or with a ligand (αGalCer or 7DW8-5, 100 ng/) in 2.5 μL at the same coordinates. The incision was closed with 2–3 interrupted 7.0 Ethicon^® sutures (Ethicon, Inc. Somerville, NJ). A subcutaneous injection of 0.03–0.1 mg/kg buprenorphine (Buprenex^® RBH, Hull, England) was given for pain control.

2.9. Bioluminescence imaging

Isofluorane anesthetized animals were imaged using the IVIS^® system (IVIS, Xenogen Corp) 10 minutes after 150 mg/kg D-luciferin (Xenogen) was injected intraperitoneally. The photons emitted from luciferase-expressing cells within the animal body and transmitted through the tissue were quantified using “Living Image”, a software program provided by the same manufacturer. A pseudo-color image representing light intensity (blue least intense and red most intense) was generated and superimposed over the grayscale reference image. Animals were imaged weekly. They were regularly examined for any neurological deficits, weight loss or signs of stress and euthanized according to pre-set criteria, in accordance with the Baylor College of Medicine’s Center for Comparative Medicine guidelines.

2.10. Statistical analysis

The statistical analysis was performed using GraphPad Prism^™ 5.0 software (GraphPad Software) or the R project (http://www.r-project.org). Comparisons between groups were based on t-test or one-way ANOVA with the Bonferroni post-test to compare selected experimental groups according to the experimental design. All statistical tests were two-sided, and a P-value < 0.05 was considered statistically significant.

3.1. CD1d is expressed in human MB cell lines and primary tumors

To examine CD1d expression in MB cells, we performed flow cytometry analysis of four established human MB cell lines and found that two of them (DAOY and MED8A) expressed CD1d on the cell surface (Fig. 1A). RT-PCR analysis confirmed CD1d expression in both DAOY and MED8A cell lines while two CD1d–negative cell lines (D341 and D283) did not express CD1d at the mRNA level (Fig. 1B). The immunohistochemical staining with an anti-CD1d mAb detected CD1d expression in nine of twenty primary MB tumors (Fig. 1C). Importantly, the majority if not all tumor cells in the CD1d-positive specimens expressed CD1d. In contrast, neurons and other cells within normal brain tissues including microglia were CD1d-negative (Suppl. Fig. 1). To examine whether the level of CD1d gene expression is associated with a known histological or molecular type of MB or with disease outcome, we performed qRT-PCR analysis of CD1d mRNA expression in 80 primary tumors that were histologically classified as classic (n = 48), desmoplastic (n = 17), and anaplastic/large cell (n = 15) types. Among these, the molecular subgroups of 61 tumors were determined (Shh, n = 25, Group 3, n = 19, and Group 4, n = 17) using a RT-PCR gene signature (manuscript in preparation). Wnt group represents less than 7% of MBs and was absent in our cohort. CD1d mRNA was expressed at similar levels across the histological types (Suppl. Fig. 2). Likewise, there was no association between the level of CD1d expression and the patient age or the disease outcome (data not shown). However, the molecular grouping did reveal that CD1d is expressed at a significantly higher level in the Shh subgroup compared to Group 4 (Fig. 1D, P = 0.015, t-test). Therefore, tumor cells in a subset of primary MB patients uniformly express CD1d on the cell surface while normal brain tissues are CD1d-negative, suggesting that CD1d could serve as a novel selective target for immunotherapy of MB.

3.2. MB cells cross-present glycolipid antigens to NKT cells in a CD1d-restricted manner

To examine whether MB cells express functional CD1d, we analyzed IFNγ and IL-4 production by NKT cells that were co-cultured with CD1d-positive MB cells in the presence or absence of a synthetic NKT ligand, αGalactosylceramide (αGalCer). Figs. 2A, B demonstrate that CD1d-positive DAOY and MED8A MB cells could effectively present αGalCer to NKT cells and induce production of IFNγ and IL-4. Pre-treatment of MB cells with anti-CD1d blocking mAb inhibited their ability to induce IFNγ in NKT cells (Fig. 2C) indicating that NKT-cell activation by MB cells is CD1d-dependent. Therefore, MB cells express CD1d without an agonistic endogenous ligand for NKT cells. However, they can effectively cross-present exogenous glycolipid antigens and activate NKT cell cytokine production in a CD1d-restricted manner.

3.3. NKT cells can specifically kill CD1d-positive MB cells

Next, we performed an in vitro cytotoxicity assay using luciferase-transduced DAOY cells as a target. Fig. 3A demonstrates that NKT cells were highly cytotoxic against MB cells pulsed with either αGalCer or its analogue 7DW8-5 [23]. At the concentration 100 ng/ml, αGalCer and 7WD8-5 were equally potent in inducing NKT cell cytotoxicity. However, titration experiments revealed that EC50 of 7WD8-5 was 10 times less than that of αGalCer (P < 0.001, Fig. 3B). We also tested NKT cell cytotoxicity after target cell pulsing with a recently discovered endogenous NKT cell ligand, β-D-glucopyranosylceramide (βGlcCer) [24]. NKT cells mediated potent cytotoxicity against βGlcCer-pulsed DAOY cells and this cytotoxicity was CD1d-dependent since it was inhibited by an anti-CD1d blocking mAb (Fig. 3C). Nearly identical results were obtained using another CD1d-positive MB cell line, MED8A while no cytotoxicity was observed against CD1d-negative MB lines (data not shown). Therefore, MB cells can cross-present exogenous and endogenous ligands and trigger potent CD1d-restricted NKT cell cytotoxcicty that could be exploited therapeutically.

3.4. Peripheral blood NKT cells are preserved in MB patients

Numerical and functional deficiencies of NKT cells have been observed in patients with some types of cancer [25;26] but not with others [20;22]. FACS analysis of primary NKT cell frequency in PBMC from 10 MB patients did not reveal a significant difference compared with NKT cell frequency in PBMC of healthy volunteers (Fig. 4A). To test whether NKT cells from MB patients can be numerically expanded ex vivo, we used αGalCer-pulsed autologous PBMC and IL-2 to induce NKT cell proliferation. Fig. 4B demonstrates that NKT cells from MB patients can be ex-vivo expanded as effectively as those from healthy people. Importantly, these patient-derived NKT cells produced high levels of IFNγ (Fig. 4C) and mediated dose-dependent cytotoxicity in response to the ligand-pulsed CD1d-positive MB cells (Fig. 4D). At the same effector to target ratios, patient-derived NKT cells were slightly less cytotoxic than those derived from healthy adults. However, the majority of NKT cells in children with MB were CD4⁺ cells (Suppl. Fig. 3), which are less cytotoxic than CD4^neg ones [27] and the latter are known to accumulate with age [28], suggesting that the observed difference in the level of NKT cell in vitro cytotoxicity is age related. Therefore, NKT cell number and functional potential are largely preserved in the peripheral blood of MB patients. These cells can be numerically expanded ex vivo and kill CD1d-positive MB cells. The lack of systemic NKT cell impairment in MB patients suggests that patient-derived NKT cells could be used for adoptive immunotherapy.

We also examined whether NKT cells are present in the MB tissues using qRT-PCR analysis of the Vα24-Jα18 invariant TCRα rearrangement in 20 primary tumors. mRNA isolated from a human NKT cell line was used as a positive control. As we have previousely reported, this method reproducibly detects one NKT cell per 10,000 tumor cells [22]. No signal for Vα24-Jα18 was detected in any of twenty tumors (data not shown). Since CD1d is recognized by both type-I and type-II NKT cells [13], all of which express CD3 and could contribute to the total T cell count, we examined the total number of tumor-infiltrating CD3+ cells in relation to CD1d expression in the same tumors. Tissue sections from 20 primary MB tumors were stained with anti-CD1d and anti-CD3 mAbs and analyzed by IHC. There was no correlation between the frequency of CD3-positive cells and the level of CD1d expression on MB cells (Suppl. Fig. 4).

3.5. Intracranial treatment with ex-vivo expanded human NKT cells has therapeutic activity in established orthotopic MB grafts in NOD/SCID mice

To test the in vivo therapeutic potential of NKT cells against MB, we used a previously described orthotopic MB model in NOD/SCID mice [12;29]. Luciferase-transduced DAOY MB cells were injected intracranially and after confirmation of tumor engraftment on day 5, mice received intracranial injections of NKT cells alone, with 7DW8-5, or equal volume of PBS (control). Tumor growth was monitored by weekly bioluminescent (BL) imaging. Fig. 5 demonstrates that NKT cells alone were able to significantly delay tumor growth compared with control (P < 0.001), and co-injection of NKT cells with 7DW8-5 resulted in even stronger anti-tumor effect and prolongation of tumor regression compared with NKT cells alone (P < 0.001) while 7DW8-5 alone had no effect (data not shown). In contrast, intravenous administration of NKT cells alone or with either 7DW8-5 or αGalCer failed to affect tumor growth. The immunohistochemical analysis of tumor tissues revealed that systemically administrated NKT cells failed to localize to the tumor site (data not shown). Therefore, the intracranial route of NKT cell administration should be considered for immunotherapy of MB patients.