Ex vivo generation of dendritic cells from cryopreserved, post-induction chemotherapy, mobilized leukapheresis from pediatric patients with medulloblastoma

Abstract

Generation of patient-derived, autologous dendritic cells (DCs) is a critical component of cancer immunotherapy with ex vivo-generated, tumor antigen-loaded DCs. An important factor in the ability to generate DCs is the potential impact of prior therapies on DC phenotype and function. We investigated the ability to generate DCs using cells harvested from pediatric patients with medulloblastoma for potential evaluation of DC-RNA based vaccination approach in this patient population. Cells harvested from medulloblastoma patient leukapheresis following induction chemotherapy and granulocyte colony stimulating factor mobilization were cryopreserved prior to use in DC generation. DCs were generated from the adherent CD14+ monocytes using standard procedures and analyzed for cell recovery, phenotype and function. To summarize, 4 out of 5 patients (80 %) had sufficient monocyte recovery to permit DC generation, and we were able to generate DCs from 3 out of these 4 patient samples (75 %). Overall, we successfully generated DCs that met phenotypic requisites for DC-based cancer therapy from3 out of 5 (60 %) patient samples andmet both phenotypic and functional requisites from 2 out of 5 (40 %) patient samples. This study highlights the potential to generate functional DCs for further clinical treatments from refractory patients that have been heavily pretreated with myelosuppressive chemotherapy. Here we demonstrate the utility of evaluating the effect of the currently employed standard-of-care therapies on the ex vivo generation of DCs for DC-based clinical studies in cancer patients.

Introduction

Malignant brain tumors represent one of the most frequent causes of cancer-related death in children. Medulloblastoma, which represents 15–20 % of all childhood brain tumors, is the most common malignant pediatric brain cancer. Despite aggressive and highly toxic multi-modality therapy including surgery, radiation therapy, and intensive chemotherapy including myeloablative regimens coupled with peripheral blood stem cell (PBSC) transplantation, from a third up to almost one-half of the children diagnosed with malignant brain tumors will still die from recurrent disease [1–7]. Furthermore, survivors often have severe lifelong treatment-associated endocrine and neuropsychologic dysfunction. The development of effective therapies that are specific and not associated with additional toxicity is critical in improving clinical outcomes for children affected by malignant brain tumors. Immunotherapy is a promising treatment strategy that could meet this clear and urgent need [8–11].

The success of immunotherapy depends on the induction of effective immune responses to treat primary tumors, and the generation of memory responses to prevent recurrence. The nature and magnitude of an immune response is determined by the way an antigen is presented to the immune system. Dendritic cells (DCs), recognized as specialized antigen presenting cells, are uniquely equipped to initiate and regulate immune responses [12], making them a key target for developing new therapies [13]. Immunizing with antigen-loaded DCs is a powerful method of inducing CD4+ and CD8+ T cells and antibodies (Abs). Our lab has developed an immunotherapy strategy that uses autologous DCs transfected with tumor antigen as cancer vaccines [14–18], that has now been validated in multiple phase I and phase II clinical trials in adults with malignant disease, including brain cancer [19, 20]. Importantly, Sipuleucel-T, an autologous cellular immunotherapy of antigen presenting cells loaded with tumor antigen, is FDAapproved for the treatment of asymptomatic or minimally symptomatic metastatic castrate resistant prostate cancer [21].

In spite of promising immunotherapy results in adults with malignant brain tumors, efforts in the immunological treatment of pediatric brain tumors have been relatively limited. This is due, at least in part, to the fact that tumor tissue available for cancer vaccine preparation is often limited in pediatric brain cancers, making the design of cancer vaccines difficult in this population. We have addressed this concern and have demonstrated that sufficient RNA for clinical vaccine preparations can be amplified with high fidelity from as few as 500 isolated tumor cells, thus allowing vaccine preparation from surgical biopsies [22]. Other equally critical factors include the feasibility of generating DCs loaded with tumor antigen ex vivo and investigating the function of peripheral blood mononuclear cells (PBMCs) (including DCs and lymphocytes) following cytotoxic therapies (radiation therapy and/or high dose chemotherapy) in children with recurrent brain tumors. In this study we determined the feasibility of DC generation from cryopreserved, post-induction chemotherapy, granulocyte colony stimulating factor (G-CSF) mobilized leukapheresis obtained from children diagnosed with medulloblastoma, for application of DC-RNA-based vaccination in this patient population.

Materials and methods

Cells

Peripheral blood mononuclear cells (PBMCs) from pediatric patients with a diagnosis of medulloblastoma (n = 5) were obtained via peripheral blood leukapheresis after induction chemotherapy and G-CSF mobilization. Briefly, patients were treated with induction chemotherapy regimens consisting of cyclophosphamide (2 gm/m²/day for 2 days every 4 weeks) followed by G-CSF mobilization until peripheral blood CD34+ stem cell (PBSC) count reached greater than 10 cells/µl. Leukapheresis was conducted according to institutional guidelines until target dose of CD34+ PBSCs were harvested (≥6 × 10⁶ cells/kg). PBMCs containing mobilized PBSCs were collected by Ficoll gradient and cryopreserved in human AB serum containing 10 % DMSO. Cryopreserved cells harvested in the years 2000–2006 in excess of that required for transplantation were obtained under Duke Institutional Review Board (IRB) approval and used in this study.

RNA

The genes encoding the full-length influenza matrix protein (Flu M1), cytomegalovirus phosphoprotein 65 (CMV pp65), human survivin and green fluorescent protein (GFP) were inserted into the pSP73-Sph/A64 plasmid using PCR and standard molecular biological techniques [23]. Plasmids were digested with SpeI for use as a template for in vitro transcription reactions using the mMESSAGE mMACHINE T7 kit (Ambion) according to the manufacturer’s protocol [24]. mRNA was purified with the RNeasy mini kit (Qiagen).

Generation and electroporation of DCs

Cells were thawed, washed in PBS and resuspended at 2 × 10⁸ cells in 30 ml AIM-V media (Invitrogen) in T-150 tissue culture flasks [25]. Cells were incubated for 1 h at 37 °C, 5 % CO₂ in a humidified incubator. The non-adherent cells were harvested by rocking the flask from side to side to dislodge them. The adherent cells were replenished with 30 ml AIM-V supplemented with 800 U/ml human GM-CSF and 500 U/ml human IL-4, then incubated at 37 °C. DCs were harvested on day 6, by collecting all non-adherent cells, followed by a cold PBS wash. Cells that were still adherent were dissociated with cell dissociation buffer (Invitrogen), 37 °C for 20 min. DCs were washed, counted and maintained on ice until use.

DCs were electroporated in 2-mm cuvettes (200 µl) at 300 V for 500 µs using an Electro Square Porator (ECM 830, BTX, San Diego, CA). Tumor antigen mRNA was used at 3 µg/10⁶ DCs [24].

RNA-electroporated DCs were matured for 8–10 h in AIM-V media containing GM-CSF (800 U/ml), IL-4 (500 U/ml) and the maturation cytokine cocktail of TNF-α (10 ng/ml), IL-1β (10 ng/ml), IL-6 (1000 U/ml), and PGE₂ (1 µg/ml) [25]. Recombinant human granulocyte macrophage colony stimulating factor (GM-CSF) was clinicalgrade Leukine (sargramostim) from Berlex Laboratories. All cytokines were obtained from R&D Systems. PGE₂ was purchased from Sigma.

Mixed lymphocyte reaction (MLR)

The allostimulatory capacity of the DCs generated from medulloblastoma patients and from adult healthy volunteer PBMCs was compared using an MLR. Allogeneic PBMCs from two healthy volunteers were co-cultured with medulloblastoma patient DCs and healthy donor DCs at a responder to stimulator ratio of 10:1. 5 × 10⁴ PBMCs were co-cultured with 5 × 10³ DCs in triplicates in 200 µl RPMI with 10 % FCS in a 96-well U-bottom plate. After 3 days of culture, 1 µCi/well of tritiated thymidine was added for 18 h and incorporation of thymidine was measured using a beta-counter. Phytohemeagglutinin (PHA) was used at a concentration of 1 µg per well.

In vitro stimulation of T cells with RNA-transfected DCs

PBMCs were thawed and resuspended in PBS and treated with DNase I (Sigma) at 200 U/ml for 20 min at 37 °C. DNase I-treated PBMCs were incubated for 1 h at 37 °C, 5 % CO₂ in a humidified incubator. Non-adherent cells were harvested and stimulated with RNA-transfected, matured DCs at a responder cell to stimulator DC ratio of 10:1 in the presence of 25 ng/ml IL-7. All stimulations were done in RPMI 1640 with 10 % FCS, 2 mM l-glutamine, 20 mM HEPES, 1 mM sodium pyruvate, 0.1 mM MEM non-essential amino acids, 100 IU/ml penicillin, 100 µg/ml streptomycin and 5 × 10⁻⁵ M β-mercaptoethanol (CTL stimulation medium). The responder T cell concentration was 2 × 10⁶ cells/ml. IL-2 was added at 100 U/ml on day 3 and every 4–5 days for 12–14 days. T cells were maintained at 1–2 × 10⁶ cells/ml in CTL stimulation medium. Alternatively, on day 7, the T cells were restimulated with RNA-electroporated DCs at a responder to stimulator ratio of 10:1 in complete RPMI-10 % FCS in the presence of 50–100 U/ml IL-2, with the T cells at 1–2 × 10⁶ cells/ml in CTL stimulation medium. T cells were harvested on day 12–14, counted and used as effector T cells in a europium-release CTL assay. Autologous DCs transfected with tumor antigen-encoding mRNA were used as targets.

In vitro cytotoxicity assay

mRNA-electroporated target cells were harvested, washed to remove all traces of media and labeled with europium (Eu). The Eu-labeling buffer (1 ml per target) contains 1 ml HEPES buffer (50 mM HEPES, 93 mM NaCl, 5 mM KCl, 2 mM MgCl₂, pH 7.4), 10 µl Eu (10 mM EuCl₃·6H₂O in 0.01 N HCl), 5 µl DTPA (100 mM diethylenetriamine pentaacetate in HEPES buffer) and 4 µl DS (1 % dextransulfate) [26]. 5 × 10⁶ target cells were resuspended in 1 ml of the europium-labeling buffer very gently and incubated on ice for 20 min. 30 µl of CaCl₂ solution (100 mM) was then added to the labeled cells, mixed and the cells were incubated for another 5 min on ice. 30 ml of Repair buffer (HEPES buffer with 10 mM glucose, 2 mM CaCl₂) was added to the cells and the cells were centrifuged at 1000 rpm for 10 min. Cells were counted and 5 × 10⁶ cells were washed 4 times with Repair buffer. After the final wash the cells were resuspended in CTL stimulation medium with no penicillin–streptomycin at 10⁵ cells/ml.

1 × 10⁴ europium-labeled targets (T) and serial dilutions of effector T cells (E) at varying E:T ratios were incubated in 200 µl of CTL stimulation medium with no penicillin– streptomycin in 96-well V-bottom plates. The plates were centrifuged at 500 × g for 3 min and incubated at 37 °C for 4 h. 50 µl of the supernatant was harvested and added to 150 µl of enhancement solution (Wallac, Perkin-Elmer) in 96-well flat-bottom plates and europium release was measured by time resolved fluorescence using the VICTOR3 Multilabel Counter (Perkin-Elmer). Specific cytotoxic activity was determined using the formula: % specific release = [(Experimental release − spontaneous release)/(Total release − spontaneous release)] × 100. Spontaneous release of the target cells was less than 25 % of total release by detergent. Spontaneous release of the target cells was determined by incubating the target cells in medium without T cells. All assays were done in triplicates.

Results

G-CSF mobilized PBMCs from five pediatric medulloblastoma patients were used in this study. Our studies include phenotypic analysis of the starting mobilized cell population, DC phenotype and evaluation of DC function. A summary of PBMC analysis and DC generation from the 5 samples is shown in Table 1. Our results indicate that using a standard, clinically applicable, monocyte-based DC generation protocol we could generate DCs from cryopreserved PBMCs obtained from medulloblastoma patients after induction chemotherapy and G-CSF mobilization.

Table 1.

Summary of dendritic cell generation from G-CSF mobilized, post-induction leukapheresis

	Patient 1	Patient 2	Patient 3	Patient 4	Patient 5
Total cells, pre-freeze	3 × 109	7.77 × 109	8.9 × 109	2.8 × 1010	1.18 × 1010
Cell number post-thaw and post-ficoll	1 × 108	5 × 109	5 × 108	3 × 108	1.25 × 109
Cells analyzed	non-adherent cells	PBMCs	PBMCs	PBMCs	PBMCs
HLA (class I)	A2+	A2−, A24+, B7+	A2−, A24+, B7+	A2−, A24+, B7−	A2−, A24−, B7−
HLA-DR (class II)	ND	70%	74%	68%	30%
CD34	ND	ND	7%	48%	0%
CD4 T cell	10%	26%	12%	0%	40%
CD8 T cell	3%	10%	5%	0%	19%
CD19 B cell	1%	8%	15%	0%	20%
CD14 monocyte	ND	38%	47%	23%	0%
CD56 NK cell	13%	10%	25%	8%	11%
DC generation	Yes	Yes	Yes	Yes	No
DC yield	3.7%	5.5%	0% (no DCs)	6.5%	NA
DC quality	Standard	Standard	Poor	Standard	NA
DC phenotype	iDC / mDC	iDC / mDC	iDC / mDC	iDC / mDC	iDC / mDC
class I	ND	16% / 29%	NA	23% / 25%	NA
class II	94% / 96%	84% / 93%	NA	99% / 99%	NA
CD11c	98% / 98%	99% / 99%	NA	98% / 100%	NA
CD80	23% / 63%	9% / 11%	NA	72% / 80%	NA
CD86	87% / 97%	92% / 96%	NA	99% / 99%	NA
CD83	6% / 50%	5% / 23%	NA	14% / 67%	NA
CD14	4% / 2%	13% / 12%	NA	2% / 3%	NA
DC function	Yes	Yes	NA	ND	NA

DC yield refers to the number of iDCs obtained from the starting plated PBMC population

NA, not applicable; ND, not determined; iDC, immature DC; mDC, mature DC

Figure 1 depicts two representative analyses of PBMCs done immediately after the cryopreserved cells were thawed. As indicated in Table 1, there were differences in cellular composition in leukapheresis obtained from patients post-induction and post-mobilization. PBMCs analyzed in Fig. 1a had no CD4+ and CD8+ T cells or B cells but had a predominance of CD34+ hematopoietic progenitor cells and CD14+ monocytes. The PBMCs shown in Fig. 1b showed presence of very few CD34+ or CD14+ cells but had a predominant population of T cells (CD4+ and CD8+), indicating a perhaps inefficient mobilization of myeloid progenitors in this particular patient. Although presence of CD34+ progenitor cells has been documented in blood post-mobilization with G-CSF, very few CD34+ cells (less than 1 % of total) were detected in this 1 patient sample (Table 1; Fig. 1b). Since our protocol utilizes adherent CD14+ monocytes as precursors for differentiation into DCs, the 4 out of the 5 samples that demonstrated the presence of significant CD14+ monocytes were used for DC generation.

Fig. 1.

Phenotypic analysis of peripheral blood cells after induction chemotherapy and G-CSF mobilization. Cryopreserved PBMCs obtained from medulloblastoma patients after induction chemotherapy and G-CSF mobilization were thawed and analyzed as indicated in the figure. Analysis of two PBMC samples is represented in a, b. Data is depicted as a 2-color or a single-color (histogram) analysis using antibodies labeled with APC (anti-CD34, -CD14, -CD8), PE (anti-CD3, -CD56, -CD19, -class II) or FITC (anti-CD4, -class I). APC allophycocyanin, PE phycoerythrin, FITC fluorescein isothiocyanate

The phenotypic characterization of DC preparations is shown in Table 1. We were able to successfully generate and phenotypically characterize DCs from 3 out of 4 cryopreserved leukapheresis. Although one sample (Patient 3) had CD14+ monocytes for generation of DCs, the DC preparation was not qualitatively or quantitatively comparable to standard DC preparations and was not used in further analysis due to limited sample availability. Of the three successful DC preparations based on phenotype, yield, and viability, two DC preparations (Table 1, Patients 1 and 2) were further analyzed for immunologic function. Functional analysis of one of the DC preparations that met the qualitative phenotypic criteria was not performed due to lack of autologous T cells available from this patient for an antigen-specific stimulation assay (Table 1, Patient 4).

In Fig. 2 we show the phenotypic characterization of monocyte-derived immature DCs and mature DCs from medulloblastoma patients post-transfection with a model antigen (CMV pp65) encoding mRNA (Fig. 2a). The DCs were phenotypically comparable to DCs generated from a healthy adult volunteer (Fig. 2b). The cells demonstrate a standard DC phenotype and were successfully matured in the presence of the maturation cytokine cocktail (IL1β, IL-6, TNF-α and PGE₂), as indicated by the increase in the levels of CD80, CD83 and CCR7 (important for DC migration). We did observe some variation in the levels of CD80 and CD83 in the different DC preparations, which is not unusual and reflects donor-to-donor variation. The cells also consistently expressed high levels of class II, a hallmark of DC phenotype (Table 1; Fig. 2). The scatter profile of electroporated immature and mature DCs is indicated in Fig. 2c and is comparable to the scatter profile of healthy donor DCs (data not shown). Finally DCs were efficiently transfected with GFP encoding RNA (Fig. 2d), using electroporation, the standard procedure for loading DCs with RNA in our clinical studies. DCs analyzed post-RNA electroporation were phenotypically comparable to DCs that were not electroporated (data not shown).

Fig. 2.

Characterization of DCs generated from post-induction, mobilized leukapheresis from patients with medulloblastoma/PNETs. DCs were generated and electroporated with RNA as described in Methods. After overnight incubation in GMCSF + IL-4 medium with or without the maturation cytokine cocktail, cells were harvested and analyzed. Phenotype of immature and mature RNA-electroporated DCs generated from medulloblastoma patient cells (a) and healthy, adult cells (b) is shown. Data is depicted as a 2-color analysis using PerCP-labeled class II antibody versus PE-labeled antibodies as indicated in the figure (peridinin chlorophyll protein complex, PerCP). Scatter profile of monocyte-derived immature and mature RNA-electroporated DCs generated from medulloblastoma patient cells (c) is shown. (d) depicts evaluation of GFP expression in DCs electroporated with GFP RNA or control RNA followed by maturation

An allogeneic MLR and an autologous antigen-specific stimulation assay was used to functionally characterize the DCs that were generated from cryopreserved PBMCs obtained from patients with medulloblastoma post-induction and post-mobilization. The allogeneic MLR was done using adult healthy donor cells as responders and using an adult healthy donor-derived DC preparation for comparison. Figure 3a demonstrates DCs from patients with medulloblastoma are comparable to normal donor DCs in their ability to stimulate an allogeneic MLR, using a proliferation assay as a read-out. Each DC preparation was tested against two allogeneic PBMCs, indicated by PBMCA and PBMC-B in the figure, and were found to stimulate allogeneic cell proliferation similarly to DCs prepared from normal healthy donor PBMC.

Fig. 3.

Functional analysis of DCs generated from post-induction, mobilized leukapheresis from patients with medulloblastoma. a Medulloblastoma patient PBMC-derived DCs were compared to DCs derived from adult healthy donor PBMCs and analyzed in a standard MLR assay as described in “Materials and Methods” section. Each bar represents average and standard deviation (SD) of triplicate samples. DCs were generated in serum-free AIM-V media supplemented with GM-CSF and IL-4 and transfected with in vitro transcribed Flu M1 RNA (b) or in vitro transcribed CMV pp65 RNA or survivin RNA (c) and used to induce cytotoxic T cells. Autologous PBMCs were stimulated in vitro with the transfected DCs as described in “Materials and Methods” section. Effector T cells were harvested as described in the protocol and assessed for their ability to lyse antigen-expressing target cells in a standard 4-h europiumrelease assay. Induction of antigen-specific CTL was measured using as targets DCs transfected with Flu M1 RNA or survivin RNA (b), and CMV pp65 RNA or survivin RNA (c). Each lytic value indicates average and SD of triplicate samples. E:T effector cell to target cell ratio; cpm, counts per minute

A stringent assessment of DC function is using DCs loaded with antigen to stimulate an antigen-specific response in autologous T cells, specifically the stimulation of tumor associated antigen-specific responses, for example survivin [27]. To this end, DCs were electroporated with either Flu matrix protein (Flu M1)-encoding RNA (Fig. 3b) and in a subsequent experiment with CMV pp65- or survivin-encoding RNA (Fig. 3c). RNA-transfected DCs were used to stimulate autologous T cells in vitro using a standard T cell stimulation protocol (described in Methods). Analysis of T cell function was performed after 1 (Fig. 3b) or 2 stimulations (Fig. 3c) with RNA-transfected DCs by measuring the cytolytic function of T cells. Separate DCs transfected with the corresponding RNA were used as target cells to measure T cell lytic activity. As shown in Fig. 3b, c, medulloblastoma patient derived DCs primed with Flu M1, CMV pp65, or survivin could effectively elicit antigen-specific T cells that effectively lysed cognate antigen-expressing targets and did not lyse non-specific antigen expressing controls.

These results indicate that adherent monocytes from post-induction, G-CSF mobilized leukaphereses from patients with medulloblastoma can be used to generate DCs that are phenotypically and functionally similar to DCs generated from adult healthy donors, providing a potential additional platform for treatment of patients with refractory disease.

Discussion

Due to successful advances in treatment of hematologic malignancies in children, malignant brain tumors now represent one of the most frequent causes of cancer-related death in children. Despite aggressive and highly toxic multi-modality therapy including surgery, craniospinal irradiation, and high-dose chemotherapy coupled with PBSC transplant, outcome for children with recurrent medulloblastoma is dismal [1–7]. Development of more effective and less toxic tumor-specific therapies is paramount in improving clinical outcomes for children affected by medulloblastoma. Immunotherapy targeting tumorspecific antigens expressed within brain tumors is a treatment modality potentially capable of meeting this clear and urgent need.

In this study, we assessed the feasibility of generating DCs from the PBMCs collected routinely from pediatric patients undergoing induction chemotherapy and PBSC mobilization for the treatment of medulloblastoma. In addition, we investigated the function of these ex vivo generated DCs post-RNA electroporation and post-maturation. Our studies demonstrate the ability to generate DCs out of severely myelodepleted and mobilized peripheral blood samples from children with medulloblastoma, that were of comparable phenotype and functional capacity to those of normal adult donors. The inability to generate DCs from cryopreserved PBMCs obtained from Patient 3 (Table 1) could potentially be explained by prior chemotherapy exposure or the status of residual malignant disease in the pediatric patient during the time of PBMC harvest. Other groups have demonstrated the correlation between lower DC yields and the presence of an untreated malignancy in pediatric patients [28, 29]. In addition, we could not assess DC generation using cells obtained from Patient 5 (Table 1), because our protocol utilizes adherent CD14+ monocytes as precursors for differentiation into DCs and there were no CD14+ monocytes in Patient 5 PBMCs.

Analysis of DC generation and DC function using cryopreserved cells derived from heavily pre-treated medulloblastoma patients has not previously been extensively investigated. There are few in vitro studies that have evaluated generation of DCs using cells from children with solid tumor malignancies, including brain tumors [28, 30]. Of note, specific to medulloblastoma, Shilyansky et al. demonstrated the ability to generate DCs from patients with medulloblastoma (number not specified) [30]. Importantly, they loaded DCs with necrotic autologous tumor and showed functional capacity of these DCs [30]. Additionally, in a clinical study of 45 children with relapsed brain tumors, of which 5 were patients with medulloblastoma, Ardon et al. evaluated the efficacy of tumor lysate-loaded DC vaccines [31]. The median overall survival (OS) in this sub-group was 5.7 months and vaccine benefit could not be evaluated. In a phase I study in 9 children with recurrent brain tumors, including 1 medulloblastoma patient, Caruso et al. evaluated the generated tumor RNA-loaded-DC vaccine; however the patient was not vaccinated due to early progressive disease [19]. In a clinical study in 15 children with relapsed brain tumors, of which 2 were patients with primitive neuro-ectodermal tumors (PNET), Geiger et al. evaluated the efficacy of tumor lysate-loaded DC vaccines and both patients had stable disease (SD) at the time of evaluation [32]. DC vaccines were evaluated in children with brain tumors in 3 additional clinical trials (n = 12), however none of these included patients with medulloblastoma [33–35].

Based on the data presented in Figs. 2 and 3, we conclude that DCs suitable for use in clinical vaccine protocols may be generated from routinely collected G-CSF mobilized leukapheresis specimens in children with medulloblastoma. Relevant to immunotherapy applications using ex vivo-generated, tumor antigen-loaded DCs [18], we demonstrate that these DCs were easily transfected with high efficiency by RNA electroporation and were capable of stimulating, both, primary tumor antigen-specific and recall memory T cell responses in vitro. In all clinical studies in pediatric brain tumor patients (described in the paragraph above), DCs were pulsed with tumor lysates or tumor RNA. In our study we have chosen to evaluate DC function by evaluating in vitro induction of CMV pp65 and survivin-specific T cell responses by RNA-transfected DCs. Both antigens were chosen based on studies that have demonstrated the presence of these antigens in medulloblastoma [36–39]. Moreover, we have recently shown that CMV is a relevant tumor antigen in patients with GBM and CMV pp65-specific T cells recognize and lyse autologous GBM tumor cells [23]. As such, DCs transfected with a mixture of known medulloblastoma antigen RNAs or total RNA isolated from autologous tumor cells are both relevant vaccine strategies for medulloblastoma.

The yield of DCs after myelosuppressive induction chemotherapy and availability of clinical specimens in excess of that required for autologous transplant, however, may be a limiting factor in the clinical utilization of these specimens. Wewould recommend the collection specifically of additional cells for DC generation during induction and mobilization protocols to ensure adequate cellular yields for cellular therapy protocols. Increased yield of cells may also be engendered through collection of PBMCs forDC generation after surgical debulking but prior to the induction of myelosuppressive adjuvant chemotherapy. We are currently examining the generation of autologous tumor RNA-pulsed DCs from children with recurrent medulloblastoma and recurrent PNETs who have undergone extensive prior cytotoxic therapy in a phase I/II clinical trial (ClinicalTrials.gov identifier: NCT01326104; PI: Mitchell) and have observed the feasibility of generating clinical-scale DC preparations from these patients (n = 13 subjects;manuscript to be published). These feasibility studies support the rationale for continued exploration of DC-based immunotherapy in children with medulloblastoma.