Dendritic Cells Enhance Polyfunctionality of Adoptively … · total resection and subsequent leukapheresis for CMV pp65-specific T cells and CMV pp65 RNA-loaded DC generation. Patients

Translational Science

Dendritic Cells Enhance Polyfunctionality ofAdoptively Transferred T Cells That TargetCytomegalovirus in GlioblastomaElizabethA. Reap1,2,CarterM. Suryadevara1,2,3, KristenA. Batich1,2,3, Luis Sanchez-Perez1,2,Gary E. Archer1,2,3, Robert J. Schmittling1,2, Pamela K. Norberg1,2, James E. Herndon II4,Patrick Healy4, Kendra L. Congdon1,2, Patrick C. Gedeon1,2, Olivia C. Campbell1,2,AdamM.Swartz1,2,3, KatherineA.Riccione1,2, JohnS.Yi5,MohammedK.Hossain-Ibrahim1,2,AnirudhSaraswathula1,2, SmitaK. Nair1,2,3,5, AnastasieM.Dunn-Pirio1,2,TaylorM. Broome1,2,Kent J.Weinhold5, Annick Desjardins1,2,6, Gordana Vlahovic1,2, Roger E. McLendon1,2,3,Allan H. Friedman1,2, Henry S. Friedman1,2, Darell D. Bigner1,2,3, Peter E. Fecci1,2,3,Duane A. Mitchell1,2,3, and John H. Sampson1,2,3

Abstract

Median survival for glioblastoma (GBM) remains <15 months.Human cytomegalovirus (CMV) antigens have been identified inGBMbut not normal brain, providing anunparalleled opportunityto subvert CMV antigens as tumor-specific immunotherapy targets.A recent trial in recurrent GBMpatients demonstrated the potentialclinical benefit of adoptive T-cell therapy (ATCT) of CMV phos-phoprotein 65 (pp65)–specific T cells. However, ex vivo analysesfrom this study found no change in the capacity of CMV pp65-specific T cells to gain multiple effector functions or polyfunction-ality, which has been associated with superior antitumor efficacy.Previous studies have shown that dendritic cells (DC) could furtherenhance tumor-specific CD8þ T-cell polyfunctionality in vivowhenadministered as a vaccine. Therefore, we hypothesized that vacci-nation with CMV pp65 RNA-loaded DCs would enhance thefrequency of polyfunctional CMV pp65-specific CD8þ T cells afterATCT.Here,we report prospective results of a pilot trial inwhich 22

patients with newly diagnosed GBM were initially enrolled, ofwhich 17 patients were randomized to receive CMV pp65-specificT cellswithCMV-DCvaccination (CMV-ATCT-DC)or saline (CMV-ATCT-saline). Patients who received CMV-ATCT-DC vaccinationexperienced a significant increase in the overall frequencies ofIFNgþ, TNFaþ, and CCL3þ polyfunctional, CMV-specific CD8þ

T cells. These increases in polyfunctional CMV-specificCD8þ T cellscorrelated (R¼ 0.7371, P¼ 0.0369)with overall survival, althoughwe cannot conclude this was causally related. Our data implicatepolyfunctional T-cell responses as a potential biomarker for effec-tive antitumor immunotherapy and support a formal assessmentofthis combination approach in a larger randomized study.

Significance: A randomized pilot trial in patients with GBMimplicates polyfunctional T-cell responses as a biomarker foreffective antitumor immunotherapy. Cancer Res; 78(1); 256–64.�2017 AACR.

IntroductionGlioblastoma (GBM) is the most common primary malignant

brain tumor and has a median survival of <15 months despite anaggressive clinical standard of care, including maximal surgicalresection, high-dose radiation, and dose-intensified temozolo-mide chemotherapy (1).Novel therapies are urgently needed, andimmunotherapy has recently emerged as a highly promisingtherapeutic approach for cancer.

We and others have previously reported the presence ofcytomegalovirus (CMV) antigens in 90% of GBMs but not innormal brain (2–4). The presence of these unique and immu-nogenic antigens presents an opportunity to leverage CMV-specific immunity against GBM while minimizing the potentialfor toxicity. In maximizing antitumor T-cell responses, it isbecoming increasingly clear that polyfunctional T cells, whichsimultaneously express more than one effector function, areproving critical for effective anticancer immunity. Recently,Crough and colleagues also demonstrated that CMV-specificT cells in patients with GBM have attenuated abilities to generatemultiple cytokines and chemokines, which is uncharacteristic of

1Duke Brain Tumor Immunotherapy Program, Department of Neurosurgery, DukeUniversity Medical Center, Durham, North Carolina. 2The Preston Robert TischBrain Tumor Center, Duke University Medical Center, Durham, North Carolina.3Department of Pathology, Duke University Medical Center, Durham, NorthCarolina. 4Department of Biostatistics and Bioinformatics, Duke University Med-ical Center, Durham, NorthCarolina. 5Division of Surgical Sciences, Department ofSurgery,DukeUniversityMedicalCenter,Durham,NorthCarolina. 6Department ofNeurology, Duke University Medical Center, Durham, North Carolina.

Note: Supplementary data for this article are available at Cancer ResearchOnline (http://cancerres.aacrjournals.org/).

D.A. Mitchell and J.H.Sampson contributed equally to this article as co-seniorauthors.

Current address for D.A. Mitchell: Preston A. Wells, Jr., Center for Brain TumorTherapy, UF Brain Tumor Immunotherapy Program, Department of Neurosur-gery, UF Medical Center, 1149 S Newell Dr, L2-100, Gainesville, FL.

Corresponding Authors: John H. Sampson, Department of Neurosurgery, DukeUniversityMedical Center, DUMCBox 3050, Durham, NC 27710. Phone: 919-684-9041; Fax: 919-684-9045; E-mail: [email protected]; and Duane A.Mitchell, [email protected]

doi: 10.1158/0008-5472.CAN-17-0469

�2017 American Association for Cancer Research.

CancerResearch

Cancer Res; 78(1) January 1, 2018256

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

CMV-specific T cells in healthy virus carriers (5). However,when cultured ex vivo with HLA-matched CMV peptides andIL2, these T cells became polyfunctional and appeared to induceantitumor immunity when transferred back into a single patientwith recurrent GBM (5). Moreover, another recent clinical trialinvestigated adoptive immunotherapy with CMV-specific T cellsin patients with recurrent GBM and showed that 11 patientsinfused with ex vivo expanded CMV-specific T cells had a prom-ising median overall survival (OS) of 13.4 months and a medianprogression-free survival (PFS) of approximately 8.1 months (6).This suggests that adoptive T-cell therapy (ATCT) may also be apromising approach for recurrent GBM (6). Importantly, howev-er, ex vivo analyses from this study found no remarkable change inthe polyfunctionality of CMV-specific T cells.

Dendritic cells (DC) are potent antigen-presenting cells, play acentral role in controlling immunity, and are among the mostfrequently used cellular adjuvants in experimental immunother-apy trials. Prior work has shown that DCs can positively impactthe polyfunctionalilty of T cells (7, 8). Moreover, a recent retro-spective study by Wimmers and colleagues suggested a linkbetween polyfunctional T-cell responses induced by DCs andlong-term tumor control in end-stage melanoma patients (9).With these studies in mind, we hypothesized that vaccinationwith CMV phosphoprotein 65 (pp65) RNA-loaded DCs wouldenhance the frequency of polyfunctional CMV-specific T cells afterATCT and therefore improve outcomes of GBM patients.

Herein, we report the safety and feasibility of using CMV pp65RNA-pulsed DCs to enhance the polyfunctionality of adoptively

transferred CMV pp65-specific T cells in a randomized pilot trial inpatients with newly diagnosed GBM. Immunotherapy targeted theimmunodominant CMV antigen pp65. Patients randomized toreceive CMV pp65-specific T cells and CMVpp65 RNA-loadedDCs(CMV-ATCT-DC) had a significant increase in the overall frequen-cies of polyfunctional CMV pp65-specific CD8þ T cells capable ofsimultaneously expressing IFNg , TNFa, and CCL3. Furthermore,within this treatment group, the increase in polyfunctional CMVpp65-specific CD8þ T-cell frequency did correlate with OS, con-firming the results found by Wimmers and colleagues in melano-ma, although we cannot conclude this was causally related.

Patients and MethodsStudy design and participants

We conducted a randomized, parallel, single-blind, single-institution pilot clinical trial at Duke University Medical Center(Durham, NC). The study schematic is summarized in Fig. 1. Thisprotocol was reviewed and approved by the FDA and the Insti-tutional Review Board at Duke University (Durham, NC). Thisstudy was conducted according to the Declaration of Helsinki,Belmont Report, U.S. Common Rule guidelines, and the Inter-national Ethical Guidelines for Biomedical Research InvolvingHuman Subjects (CIOMS). All patients signed a written informedconsent before study inclusion.

This trial recruited 22 CMV-seropositive patients with con-firmedWorldHealthOrganization grade IVGBMand aKarnofskyperformance scale (KPS) score �80. Patients underwent gross

Figure 1.

Trial design. As per the clinical standard of care, patients underwent surgical resection and received xRT with concurrent temozolomide (75 mg/m2) over a 6-weekperiod. Three to 4 weeks after xRT/TMZ, patients received cycle 1 of a temozolomide (200 mg/m2/day) daily for 5 days starting on day 1. Leukapheresiswas performed prior to chemoradiation. DC generation began at leukapheresis, followed by pp65mRNA transfection on day 7 of culture. Expansion of T cells beganon day 8, ending on day 35 to 42. On day 21, patients received CMV pp65-specific T cells with either CMV pp65 RNA-loaded DCs or saline by random assignment.T-cell activity was measured before vaccine 1 on day 21 and 7 days after vaccine 1 at day 28. A total of three CMV pp65-loaded DC or saline infusions wasadministered in 14-day intervals. All patients underwent leukapheresis after day 49 and then received a cycle of temozolomide every 28 days for at least 6 cycles.Imaging was performed bimonthly, and upon tumor progression, suitable participants underwent stereotactic biopsy or resection as standard of care.Patients were followed until death.

CMV DC Vaccines Enhance Polyfunctionality of ATCT in Glioma

www.aacrjournals.org Cancer Res; 78(1) January 1, 2018 257

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

total resection and subsequent leukapheresis for CMV pp65-specific T cells and CMV pp65 RNA-loaded DC generation.Patients then received conformal external beam radiotherapywith concurrent temozolomide (xRT/TMZ) (75 mg/m2/day)over a 6-week timeframe (Fig. 1). Temozolomide was discon-tinued in one patient due to intolerability. Twenty-one to 28days after xRT/TMZ, patients received a 5-day cycle of temo-zolomide at a dose of 200 mg/m2/day as per standard of care.After completion, patients were screened by CT or MRI forevidence of progressive disease, and 5 patients were excludedprior to immunotherapy initiation on this basis. Patients wereregularly monitored for clinical decline or progression and wereremoved from study if either occurred before experimentaltherapy was complete. Two patients began experimental treat-ment before progression was confirmed radiographically andwere immediately removed from study and excluded fromimmune monitoring and clinical analyses, but were includedin safety analyses. A total of 17 patients were randomlyassigned to receive CMV pp65-specific T cells with either CMVpp65 RNA-loaded DC (n ¼ 9) or saline (n ¼ 8), and 15 patientsreceived all 3 vaccinations (Supplementary Fig. S1).

In an attempt to treat these tumors by targeting CMV anti-gens, we conducted an ATCT of ex vivo expanded CMV pp65-specific T cells and then randomized patients to receive eitherDCs loaded with pp65 RNA (n ¼ 8) or saline (n ¼ 7) todetermine whether a DC vaccine could enhance the functionof the transferred T cells. Patients received three CMV pp65RNA-loaded DC or saline infusions in 14-day intervals. Fol-lowing the third administration, all patients underwent leuka-pheresis or blood draw and later received 5-day cycles oftemozolomide every month (150–200 mg/m2/day) for 6 to25 cycles. In certain cases, temozolomide dosing was modifiedduring maintenance cycles as determined by the attendingphysician. MRI was performed bimonthly to monitor disease,and upon progression, suitable participants underwent stereo-tactic biopsy or resection as standard of care. Radiographicprogression was defined by the Response Assessment in Neuro-Oncology (RANO) criteria. Clinical progression was defined bysignificant change in overall neurologic status, a change in KPSof �30 points, or the development of a new focal neurologicdeficit. The study schematic is summarized in Fig. 1. Baselinepatient characteristics are summarized in Table 1. Adverse

events with possible relation to experimental treatment aresummarized in Supplementary Table S1.

In vitro generation of CMV pp65-specific T cells and CMV pp65-loaded DCs

To generate CMV pp65-specific T cells and CMV pp65 RNA-loaded DCs, peripheral blood mononuclear cells (PBMC) wereobtained by leukapheresis. Monocyte precursors for DC genera-tionwere isolated by plastic adherence inAIMVmedia containing2% human AB serum (HABS) for 1 hour at 37

�C, 5%CO2. After 1

hour, nonadherent cells were collected and cryopreserved to beused for CMV pp65-specific T-cell generation. The adherent cellswere incubated for 7 days at 37

�C, 5% CO2 in AIM V media

containing 800units ofGM-CSF and500units of IL4permL.After7 days, immature DCs were harvested and transfected with CMVpp65-mRNA by electroporation. The transfected DCs were cul-tured with a maturation cocktail of TNFa (10 ng/mL), IL1b(10 ng/mL), and IL6 (1,000 U/mL) in AIM V containing GM-CSF and IL4 for 18 to 24 hours at 37

�C, 5% CO2. DCs were

harvested and cryopreserved in 80%HABS, 10%DMSO, and 10%dextrose. Patients randomized to the CMV-ATCT-DC cohortreceived three separate infusions of 2 � 107 DCs. according tothe treatment schedule shown in Fig. 1.

For CMV pp65-specific T-cell generation, the previously cryo-preserved nonadherent (NA) cells were thawed, and cells weremixed at a ratio of 1:10 pp65-loaded DCs:NAs in AIM V mediawith 2% HABS, and IL2 was added at a concentration of100 IU/mL on day 3. Cells were adjusted to 1–5� 106/mL duringexpansionwith freshmedia and IL2.Cells were cryopreserved aftera dose of 3 � 107 was achieved. Prior to infusion, CMV pp65-specific T cells were thawed and resuspended in saline containing1% human serum albumin at a concentration of 2.5 � 107

cells/mL. Patients received an intended dose of 3 � 107 CMVpp65-specific T cells/kg. CMV pp65-specific T-cell responses weremeasured by direct ex vivo IFNg ELISPOT assay as described pre-viously (10).Resultswere expressedas themeanspot-formingcells/106 PBMCs after subtraction of counts from cells cultured with nopeptide. Polychromatic flow cytometry was used to identify cellpopulationspresent inpre- andpostexpandedCMVpp65-specific Tcells prior to CMV pp65-specific T-cell infusion. Approximately300,000 total events were collected per sample. Lymphocyte-gatedevents ranged between 75,000 and 250,000 events in data shown.

Table 1. Baseline characteristics of patients included in the immune monitoring and efficacy analyses

Patient ID Age (y) Race Sex MGMT promoter status Random assignment KPS CMV serostatus

ER1 35 C M Unmethylated CMV-ATCT-Saline 90 þER2 71 C F Unmethylated CMV-ATCT-DC 80 þER3 68 C M Unmethylated CMV-ATCT-Saline 90 þER4 50 C M Methylated CMV-ATCT-DC 90 þER5 54 C M Methylated CMV-ATCT-DC 90 þER6 73 C F Unmethylated CMV-ATCT-Saline 90 þER7 46 C F Not done CMV-ATCT-Saline 90 þER8 73 C M Unmethylated CMV-ATCT-DC 90 þER9 57 C M Unmethylated CMV-ATCT-DC 90 þER10 47 C M Unmethylated CMV-ATCT-DC 90 þER11 43 C M Unmethylated CMV-ATCT-Saline 90 þER12 61 AA M Unmethylated CMV-ATCT-DC 80 þER13 51 C M Unmethylated CMV-ATCT-Saline 80 þER14 55 C F Unmethylated CMV-ATCT-DC 80 þER15 59 C M Unmethylated CMV-ATCT-Saline 90 þNOTE: Age was recorded at the time of patient consent.Abbreviations: AA, African American; C, Caucasian; ER, ERaDICATe; MGMT, O-6-methylguanine DNA methyltransferase.

Reap et al.

Cancer Res; 78(1) January 1, 2018 Cancer Research258

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

Immune monitoring and T-cell profilingWe collected blood samples from patients before and 7 days

after CMV pp65-specific T-cell infusion and the first CMV pp65RNA-loaded DC or saline administration. PBMCs were separatedwithin 6 hours of collection by density centrifugation on Histo-paque (Sigma 1077), frozen to �80�C at a rate of 1�C/minute,and stored in liquid nitrogen. On day of testing, PMBCs werethawed at 37�C, washed, resuspended in R-10 medium contain-ing serum, and cell number and viabilityweremeasured byGuavaCounter (Millipore). All samples recorded a viability >75% afterthawing. CMV pp65-specificity of T cells was confirmed by CMVpp65-specific IFNg ELISPOT assay. Polyfunctionality of CMVpp65-specific CD8þ T cells was evaluated using polychromaticintracellular flow cytometry as described previously (10). Thefrequency of antigen-specific CD8þ T cells producing one, two,and/or three cytokines was calculated using FlowJo software andanalyzed. The data analysis software Simplified Presentation ofIncredibly Complex Evaluations (SPICE) was used to analyzeand produce representations of high content and multivariatedatasets.

Statistical analysesOSwas computed from the date of surgical resection to the date

of death. All patients were followed until death. Survival distribu-tions are described using Kaplan–Meier methods, and associa-tions of survival outcomes with polyfunctional T-cell frequencieswere assessed using the Pearson correlation coefficient. The log-rank test was used to compare survival distributions between thetwo treatment groups. Wilcoxon signed rank tests were used toassess changes in cell frequencies before and after immunother-apy. A Kruskal–Wallis nonparametric ANOVA followed by Dunnpairwise comparison was used to assess differences in IFNgmedian fluorescence intensity (MFI) among functional subsetsof CMV pp65-specific CD8þ T cells.

As an exploratory effort to assess the impact of CMVpp65 RNA-pulsedDCs in enhancing T cells, this pilot study was not designedto detect clinically important differences between randomizedgroups with reasonable power. Without adequate power, a

hypothesis test lacking statistical significance does not eliminateor minimize the possibility that a clinically important differencemay truly exist.

ResultsPatient characteristics and safety

Twenty-two CMV-seropositive patients with newly diagnosedGBMwere consented into the clinical trial Evaluation of RecoveryFrom Drug-Induced Lymphopenia Using Cytomegalovirus-spe-cific T-cell Adoptive Transfer (ERaDICATe) after undergoing sur-gery. All of these patients met protocol eligibility with residualradiographic contrast enhancement on postresection computer-ized axial tomography (CT) orMRI of <1 cm inmaximal diameterin any axial plane. Leukapheresis was performed prior to stan-dard-of-care chemoradiation (see Materials and Methods) asshown in the accompanying trial schema (Fig. 1) and patientflow diagram (Supplementary Fig. S1). Five of the 22 consentedpatients exhibited clinical decline or progressive disease during orafter completion of chemoradiation and were removed fromstudy evaluations before randomization. The remaining 17patients were randomly assigned to receive either CMV-ATCT-DC(n¼9)orCMV-ATCT-saline (n¼8). Twoof these 17patients (1 ineach arm) initiated treatment even though prevaccination MRIfindings were inconclusive and did not definitively rule outdisease progression. In these 2 patients, follow-up MRI wasconducted one month later. At that time, progression was con-firmed, and these 2 patients were withdrawn from study partic-ipation. During this month, the patient in the CMV-ATCT-salinearm received one treatment, and the patient in theCMV-ATCT-DCarm received two vaccinations. These patients were not includedin efficacy and immune monitoring analyses. Hence, efficacy andimmunemonitoring analyses were limited to the 15 patients whocompleted treatment while safety analyses included all 17patients.

Baseline demographics and patient characteristics were com-parable between treatment arms (Table 1). The first patientwas consented on August 1, 2008, and the last patient expired

Figure 2.

Characterization and responsiveness ofin vitro expanded CMV pp65-specific Tcells. Patient PBMCs were coculturedwith autologous DCs loaded with CMVpp65-encoding RNA and expanded invitro in thepresence of IL2.A,Peripheralblood lymphocytes were phenotypedbefore and after in vitro expansion.B, Magnitude of CMV pp65-specificT-cell responses was measured beforeand after expansion by CMV pp65 IFNgELISpot assay after stimulation. SFC,spot-forming cell. Because ofinsufficient numbersof cells from2of 15,13 patients' cells were analyzed forphenotype and function. Statisticalsignificance was determined byWilcoxon signed rank test.

CMV DC Vaccines Enhance Polyfunctionality of ATCT in Glioma

www.aacrjournals.org Cancer Res; 78(1) January 1, 2018 259

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

on April 7, 2015. The trial was ended after the first cohort of atleast 12 patients successfully completed the trial.

In general, immunotherapy was safe, well tolerated, and pro-duced only minor adverse events (AE) consistent with thoseexpected in this patient population following the clinical standardof care. Toxicity grading was assigned according to the NCICommon Terminology Criteria for Adverse Events (Version3.0). No severe AEs were produced by treatment. Patients weremonitored by the attending physician andmedical staff and weremanaged with routine clinical practice as required. AEs are sum-marized in Supplementary Table S1.

In vitro expansion of patient PBMCs enriches for CMV pp65-specific CD8þ T cells

Adoptive T-cell immunotherapy and DC vaccination weredesigned to target the immunodominant CMV antigen pp65. Weestablished and qualified a protocol (see Materials and Methods)to selectively expand CMV pp65-specific T cells and DCs loadedwith pp65 RNA from patient-derived PBMCs to clinical scale.Both cellular products were generated prior to planned random-ization for all patients initially recruited for study and weremanufactured with sufficient quantity and quality tomeet releasecriteria in 100% of patients (n ¼ 22).

In vitro culturing conditions expanded the mean frequency ofCD3þ T cells from 77.42% to 97.27% of total cells (Fig. 2A, P ¼0.0002). The mean frequency of CD4þ T cells was reduced from53.11% to 19.79% (Fig. 2A, P ¼ 0.0005), while CD8þ T cellscomprised 72.16% of the postexpansion product comparedwith just 39.04% before expansion (Fig. 2A, P ¼ 0.0010). Thefinal infusion products were characterized prior to infusion; themean frequencies of CD8þ T cells, CD4þ T cells, regulatory Tcells, B cells, and NK cells were 70.36%, 19.09%, 6.23%, 0.17%,and 0.85%, respectively. Importantly, the expanded lympho-cyte fraction was functionally responsive to CMV pp65 antigenstimulation as measured by IFNg ELISPOT assay (Fig. 2B, P ¼0.0002). All patients, with the exception of two, received asingle infusion of 3 � 107 cells/kg as in vitro generated CMVpp65-specific T cells. These 2 patients received 71% and 72% ofthe intended doses as shown in Supplementary Table S2.Separately, CMV pp65 RNA-loaded DCs generated from autol-ogous DCs transfected with pp65-lysosomal associated mem-brane protein mRNA were matured and infused intradermallyat a dose of 2 � 107 cells in patients randomized to this arm(CMV-ATCT-DC).

CMVpp65RNA-loadedDCs enhance T-cell polyfunctionality ofCMV pp65 T cells

To assess whether vaccination with CMV pp65 RNA-loadedDCs impacted T-cell responses in peripheral blood, we performedex vivo analyses on circulating CMV pp65-specific CD8þ T cellsbefore and 7 days after patients were administered CMV-ATCT-saline or CMV-ATCT-DC. We found that patients who receivedsaline following CMV pp65-specific T cells had no significantchange in the frequency of cells positive for one marker (Supple-mentary Fig. S2A, S2C, andS2E,P¼1.0000,NS),whereas patientsin the CMV-ATCT-DC cohort exhibited significantly enhancedfrequencies of CMV pp65-specific T cells single positive for IFNg ,TNFa, or CCL3 (Supplementary Fig. S2B, S2D, and S2F, P ¼0.0078).

Next, we determined whether CMV-ATCT-DC elicited achange in the frequency of circulating polyfunctional T cells,

which are T cells that are capable of simultaneously generatingIFNg , TNFa, and CCL3 at the single-cell level. Patients in thecohort that received saline with ex vivo expanded CMV pp65-specific T cells (CMV-ATCT-saline) had no increase in IFNgþ

TNFaþ CCL3þ triple-positive CMV pp65-specific T cells(Fig. 3A, P ¼ 1.000 NS), while patients randomized to receiveCMV pp65-specific T cells and CMV pp65 RNA-loaded DCs(CMV-ATCT-DC) had a significant increase in the mean fre-quency of IFNgþ TNFaþ CCL3þ CMV pp65-specific CD8þ Tcells (Fig. 3B, P ¼ 0.0078). To confirm detection of polyfunc-tional CMV-specific T cells, we compared the MFI of IFNgexpression in T cells defined by 3, 2, and 1 function(s) in avalidated assay as described previously (10). As expected, CMVpp65-specific T cells defined by 3 functions exhibited higherMFI of IFNg than T cells with fewer functions (Fig. 3C and D,P ¼ 0.0001), demonstrating a hierarchy of IFNg expressionbetween functionally discrete subsets of T cells. This clearlyauthenticates the presence of bona fide polyfunctionality andsubstantiates previous reports that T cells with these multiplefunctions are the most potent effectors. These data collectively

Figure 3.

Assessment of polyfunctionality in circulating CMV pp65-specific CD8þ T cellsbefore and after immunotherapy. Ex vivo analysis of CMV pp65-specific CD8þ Tcells in peripheral blood circulation was performed to measure the frequency ofcells simultaneously expressing IFNg , TNFa, and CCL3 before and afterimmunotherapy with CMV-ATCT-saline (A; P¼ 1.0000, nonsignificant), n¼ 7 orCMV-ATCT-DC (B;P¼0.0078), n¼8. Statistical significancewasdeterminedbyWilcoxon signed rank test. C and D, To confirm the presence of bona fidepolyfunctionality, the MFI of IFNg expression was compared between CMVpp65-specific CD8þ T cells with three functions versus T cells with lesserfunctions. Statistical significance was determined by Kruskal–Wallisnonparametric ANOVA followed by Dunn pairwise comparison.

Reap et al.

Cancer Res; 78(1) January 1, 2018 Cancer Research260

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

support a role for CMV pp65 RNA-loaded DC vaccines in the invivo expansion and differentiation of polyfunctional CMVpp65-specific T cells.

Finally, we performed polyfunctional flow cytometry aggregatedata analyses to catalog the proportion of CMV pp65-specificCD8þ T cells defined by 3, 2, or 1 function(s) to determine howthe distribution of these individual populations may have chan-ged relative to one another in response to immunotherapy.Patients in the CMV-ATCT-saline cohort had no major changein any defined T-cell subsets, whereas patients in the CMV-ATCT-DC cohort experienced enhancements in the mean frequencies ofCMVpp65-specific T cells displaying 3, 2, or 1 function(s) (Fig. 4Aand B).

We also assessed T-cell quality before and after immunotherapyindividually in all patients (Supplementary Fig. S3) and highlight2 patients representative of the CMV-ATCT-saline (Fig. 4C) orCMV-ATCT-DC (Fig. 4D) cohort. As shown, ERaDICate (ER)patient 11 (CMV-ATCT-saline) had no major change in T-cellfunctionality after immunotherapy, whereas patient ER14(CMV-ATCT-DC) exhibited amarked reduction in the proportionof T cells monofunctional for IFNg or TNFa, with a concomitantboost in T cells displaying all three functions. These observationsare consistent with the cohorts at large; 4 of 8 patients (ER2, ER9,ER12, and ER14) receiving CMV-ATCT-DC had a dramatic reduc-tion in theproportionofmonofunctional IFNg-secreting cells thatwas accompanied by an increased proportion of T cells with all

three functions (Supplementary Fig. S3). Only one patient ran-domized to the saline arm experienced a similar shift (ER15),whereas 6 of 7 patients had little to no shift in the proportion ofmonofunctional IFNg-secreting cells following therapy. Impor-tantly, these data collectively show that polyfunctional T cellsoccupy a significantly increased fraction within the CMV pp65-specific CD8þ T-cell compartment following immunotherapywith CMV-ATCT-DC, indicating that in vivo antigenic stimulationusing DCs may be an opportune strategy to favorably shift thefunctional distribution of tumor-specific T cells to more powerfuleffector T cells. Notably, patients who received CMV-ATCT-DCalso had a significantly higher increase in polyfunctional T cells atday 7 comparedwithday 1 reflected by amean fold change of 2.48(Fig. 5, P ¼ 0.0401), whereas patients who received CMV-ATCT-saline only had a modest mean fold change in polyfunctionalT cells of 1.25.

Polyfunctional T cells correlate with outcome in vaccinatedpatients

As polyfunctional T-cell responses are considered a biomarkerof protective immunity in acute and chronic viral infections(11–13), we sought to establish their significance in this cancerstudy. Unlike the negative correlation between the fold change ofpolyfunctional T cells and improved survival in patients whoreceived CMV-ATCT-saline (Fig. 6A, R ¼ �0.4835, P ¼ 0.2716),we observed a positive correlation between the fold change of

Figure 4.

CMV pp65-specific CD8þ T cellscatalogued by quality before and afterimmunotherapy. The CMV pp65-specific CD8þ T-cell response iscomposed of distinct responders thatvary by functionality. Aggregate dataanalyseswereperformed todeterminethe relative mean distribution of CMVpp65-specific CD8þ T cells expressingone ormore functions defined by IFNg ,TNFa, and CCL3 before (gray bar) andafter (black bar) immunotherapy in allpatients who received CMV-ATCT-saline, n ¼ 7 (A) or CMV-ATCT-DC,n ¼ 8 (B). Bars represent meanfrequencies of CMV pp65-specificCD8þ T cells expressing the particularcombination of functions shown.C andD, The qualitative distribution of CMVpp65-specific CD8þ T cells before andafter immunotherapy is shown for asingle patient representative of eitherarm. ER11 and ER14 refer to theERaDICATe trial patient numbers.

CMV DC Vaccines Enhance Polyfunctionality of ATCT in Glioma

www.aacrjournals.org Cancer Res; 78(1) January 1, 2018 261

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

polyfunctional T cells and improved survival in patients whoreceived CMV-ATCT-DC (Fig. 6B, R ¼ 0.7371, P ¼ 0.0369).However, this study was not powered to detect differencesbetween cohorts with regard to the clinical outcomes of PFS andOS, so no conclusions can be drawn with regard to the causalassociation between polyfunctional T-cell responses and out-come. To our knowledge, this is the first report to show that DCvaccination positively impacts T-cell responses in peripheralblood after CMV-specific T-cell therapy and suggests furtherstudies should examine the clinical significance of enhancingT-cell polyfunctionality in immunotherapy targeting newly diag-nosed GBM.

DiscussionHere, we report our ability to produce and safely administer

CMV pp65-specific T cells with DC vaccination (CMV-ATCT-DC)

or saline (CMV-ATCT-saline) to patients with newly diagnosedGBM. The data reported here suggest important principles thatwarrant a formal assessment within a larger cohort of patients in afollow-up randomized phase II study.

In designing this study, we reasoned that an effective ATCTstrategy using CMV pp65-specific T cells would require a quali-tative shift of T cells defined by 0–1 function to >1 function in vivo,and we believed this could be achieved by an accompanying DCvaccine based on the success of this combinatorial approach instudies targeting metastatic melanoma and B-cell malignancies(9, 14). Here, polyfunctional cells were defined by their ability tosimultaneously secrete IFNg , TNFa, and CCL3. We chose toinclude CCL3 in our polyfunctional T-cell panel, as we recentlyfound CCL3 to be a critical mediator of antitumor immunother-apy in the context of tumor antigen–specific DC vaccination (15).This study demonstrates that patients with newly diagnosedGBMwho are administered CMV-ATCT-DC had significant increases inthe overall frequencies of polyfunctional CMV pp65-specificCD8þ T cells, and a greater proportion of these cells were iden-tified to harbor multiple effector functions, compared with thosereceiving CMV-ATCT-saline.

Importantly, we observed that augmented polyfunctional T-cell presence was detected in patients treated with CMV pp65RNA-pulsed DCs. These observations suggest that CMV pp65specific T-cell polyfunctionality may potentially serve as a markerof effective therapeutic response and strengthen support for therole of polyfunctional T cells in the immune system's effort againstGBMand other cancers, in addition to the already established roleof polyfunctional T cells in protective immunity against acute andchronic viral infections (11, 12). Our findings are consistent withprevious data, which described CMV-specific T cells isolated fromGBM patients as being deficient in polyfunctionality, but follow-ing antigen exposure ex vivo, had restoration of their ability togenerate multiple cytokines and were capable of mounting aneffective antitumor response in vivo (5). Also analogous to pre-vious reports ofGBMpatients receivingCMVpp65-specific T cells,patients in our study who were administered CMV-ATCT-salinedid not experience significant shifts in T-cell functionality (6).Incidentally, we were able to obtain samples either at or close toprogression on 2 patients after 9 cycles of temozolomide andfound that the polyfunctional CD8þ pp65 antigen-specific cyto-kine response was greatly diminished after temozolomide wasrestarted in one patient and increased in the other. To our

Figure 5.

Assessment of polyfunctionality in circulating CMV pp65-specific CD8þ T cellsbefore and after immunotherapy. Ex vivo analysis of CMV pp65-specific CD8þ Tcells in peripheral blood circulation was performed to measure the frequency ofcells simultaneously expressing IFNg , TNFa, and CCL3 molecules before andafter immunotherapy with CMV-ATCT-saline (white circles) or CMV-ATCT-DC(black circles). Statistical significance was determined byWilcoxon signed ranktest. The fold change of CMVpp65-specificCD8þT cellswith three functionswasdetermined for all 15 patients (P ¼ 0.0401). Statistical significance wasdetermined by the Mann–Whitney test.

Figure 6.

Correlation of fold change of polyfunctionalCMV pp65-specific CD8þ T cells with OS. Linearregression analysis was performed with the foldchange of CMV pp65-specific CD8þ T cells withsimultaneous expression of IFNg , TNFa, andCCL3 and OS for all patients randomized toCMV-ATCT-saline (A; R ¼ �0.4835, P ¼ 0.2716nonsignificant) or CMV-ATCT-DC (B;R¼0.7371,P¼0.0369).n¼ 7 forA;n¼8 forB. Associationsof survival outcomes with polyfunctional T-cellfrequencies were assessed using the Pearsoncorrelation coefficient.

Reap et al.

Cancer Res; 78(1) January 1, 2018 Cancer Research262

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

knowledge, we are the first to report that DC vaccination enhancespolyfunctionality of transferred CMV pp65-specific CD8þ T cellsin vivo in patients with newly diagnosed GBM.

In summary, the findings highlighted here have demonstratedthe safety and feasibility of adding CMV pp65 RNA-loaded-DC toCMV pp65-specific T cells in conjunction with the clinical stan-dardof care for patientswithnewly diagnosedGBM.Our data alsoreaffirm earlier experiences targeting CMV pp65 in newly diag-nosed GBM patients by identifying CMV pp65 as a robust anti-GBM antigen. It also provides evidence that CMV pp65-specific Tcells and CMV pp65 RNA-loaded DCs can be successfully gener-ated from GBM patients, who in general possess cell-mediatedimmune deficiencies. As the FDA required the testing of a singleantigen at a time for this immunotherapy trial, next steps wouldbe to employ several more antigens in a multiantigenic vaccinethat could elicit several different T cell–responding populations.Finally, these results indicate that adjuvant DC vaccinationenhances transferred T-cell immune responses in vivo and suggestthat ex vivo analyses of T-cell polyfunctionality may equip clin-icianswith a unique opportunity to help predict patient responsesfollowing immunotherapy, which is requisite in this era of indi-vidualized antitumor therapy.

Disclosure of Potential Conflicts of InterestD.A. Mitchell reports receiving a commercial research grant from Immu-

nomic Therapeutics, Inc. and has provided expert testimony for Annias Immu-notherapeutics, Inc. J.H. Sampson is a consultant/advisory board member forAnnias Therapeutics. G.E. Archer and E.A. Reap are stockholders of AnniasTherapeutics. No potential conflicts of interest were disclosed by the otherauthors.

Authors' ContributionsConception and design: E.A. Reap, G.E. Archer, P.K. Norberg, H.S. Friedman,D.A. Mitchell, J.H. SampsonDevelopment of methodology: E.A. Reap, G.E. Archer, R.J. Schmittling,P.K. Norberg, S.K. Nair, K.J. Weinhold, R.E. McLendon, D.A. Mitchell,J.H. SampsonAcquisition of data (provided animals, acquired and managed patients,provided facilities, etc.): E.A. Reap, K.A. Batich, G.E. Archer, R.J. Schmittling,P.K. Norberg, A. Saraswathula, A. Desjardins, R.E. McLendon, A.H. Friedman,H.S. Friedman, P.E. Fecci, D.A. Mitchell, J.H. SampsonAnalysis and interpretation of data (e.g., statistical analysis, biostatistics,computational analysis): E.A. Reap, C.M. Suryadevara, K.A. Batich, L. Sanchez-

Perez, R.J. Schmittling, P.K. Norberg, J.E. Herndon II, P. Healy, K.L. Congdon,P.C. Gedeon, O.C. Campbell, K.A. Riccione, J.S. Yi, M.K. Hossain-Ibrahim,S.K. Nair, K.J. Weinhold, A. Desjardins, R.E. McLendon, H.S. Friedman,D.A. Mitchell, J.H. SampsonWriting, review, and/or revision of the manuscript: E.A. Reap, C.M. Suryade-vara, K.A. Batich, L. Sanchez-Perez, G.E. Archer, R.J. Schmittling, P.K. Norberg,J.E. Herndon II, P. Healy, K.L. Congdon, P.C. Gedeon, O.C. Campbell,A.M. Swartz, K.A. Riccione, J.S. Yi, M.K. Hossain-Ibrahim, A. Saraswathula,A.M. Dunn-Pirio, T.M. Broome, K.J. Weinhold, A. Desjardins, G. Vlahovic,R.E. McLendon, H.S. Friedman, P.E. Fecci, D.A. Mitchell, J.H. SampsonAdministrative, technical, or material support (i.e., reporting or organizingdata, constructing databases): E.A. Reap, C.M. Suryadevara, K.A. Batich,G.E. Archer, P.K. Norberg, P. Healy, K.J. Weinhold, J.H. SampsonStudy supervision: E.A. Reap, G.E. Archer, D.D. Bigner, D.A. Mitchell,J.H. Sampson

AcknowledgmentsThe authors thank the medical staff, including S. Norman, B. Perry, and

D. Lally-Goss, who supported this clinical study and provided unparalleled careand comfort to our patients. We also thank S. Janetzki (Zellnet Consulting, Inc,Fort Lee, NJ, and the Cancer Immunotherapy Consortium) for evaluatingELISpot plates and M. Roederer (NIH, Bethesda, MD) for providing access toSPICE software.Wewould also like to acknowledge each of the patients we haveencountered who offer an endless source of inspiration. This trial is registeredwith www.clinicaltrials.gov (NCT00693095) as Evaluation of Recovery FromDrug-Induced Lymphopenia Using Cytomegalovirus-specific T-cell AdoptiveTransfer (ERaDICATe). This work was supported by grants from the NIHNational Institute of Neurological Disorders and Stroke (R01-NS06703 toD.A. Mitchell), NCI (R01-CA134844 to D.A. Mitchell), and the Departmentof Defense (W81XWH-10-1-0089 to D.A. Mitchell). Additional support wasprovided by the National Brain Tumor Society (D.A. Mitchell and J.H. Samp-son), the American Brain Tumor Association (D.A. Mitchell and J.H. Sampson),Accelerate Brain Cancer Cure Foundation Young Investigator's Award(D.A. Mitchell), The Kinetics Foundation (J.H. Sampson), Ben and CatherineIvy Foundation (J.H. Sampson), and in part by Duke University's Clinical &Translational Science Awards grant 1UL2 RR024128-01 from the NIHNationalCenter for Research Resources.

The costs of publication of this article were defrayed in part by the paymentof page charges. This article must therefore be hereby marked advertisement inaccordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received February 20, 2017; revised June 27, 2017; accepted October 26,2017; published OnlineFirst November 1, 2017.

References1. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ,

et al. Radiotherapy plus concomitant and adjuvant temozolomide forglioblastoma. N Engl J Med 2005;352:987–96

2. Cobbs CS, Harkins L, Samanta M, Gillespie GY, Bharara S, King PH, et al.Human cytomegalovirus infection and expression in human malignantglioma. Cancer Res 2002;62:3347–50

3. Mitchell DA, Xie W, Schmittling R, Learn C, Friedman A, McLendon RE,et al. Sensitive detection of human cytomegalovirus in tumors and periph-eral blood of patients diagnosed with glioblastoma. Neuro Oncol2008;10:10–8

4. Dziurzynski K, Chang SM,Heimberger AB, Kalejta RF,McGregor Dallas SR,Smit M, et al. Consensus on the role of human cytomegalovirus inglioblastoma. Neuro Oncol 2012;14:246–55

5. Crough T, Beagley L, Smith C, Jones L, Walker DG, Khanna R. Ex vivofunctional analysis, expansion and adoptive transfer of cytomegalovirus-specific T-cells in patients with glioblastoma multiforme. Immunol CellBiol 2012;90:872–80

6. Schuessler A, Smith C, Beagley L, Boyle GM, Rehan S, Matthews K, et al.Autologous T-cell therapy for cytomegalovirus as a consolidative treatmentfor recurrent glioblastoma. Cancer Res 2014;74:3466–76

7. Colleton BA, Huang XL, Melhem NM, Fan Z, Borowski L, RappoccioloG, et al. Primary human immunodeficiency virus type 1-specific CD8þT-cell responses induced by myeloid dendritic cells. J Virol 2009;83:6288–99

8. Van Gulck E, Vlieghe E, Vekemans M, Van Tendeloo VF, Van De VeldeA, Smits E, et al. mRNA-based dendritic cell vaccination inducespotent antiviral T-cell responses in HIV-1-infected patients. AIDS2012;26:F1–12

9. Wimmers F, Aarntzen EH, Duiveman-deBoer T, Figdor CG, Jacobs JF, Tel J,et al. Long-lasting multifunctional CD8þ T cell responses in end-stagemelanoma patients can be induced by dendritic cell vaccination. Oncoim-munology 2016;5:e1067745

10. Bernstein DI, Reap EA, Katen K, Watson A, Smith K, Norberg P, et al.Randomized, double-blind, Phase 1 trial of an alphavirus replicon vaccinefor cytomegalovirus in CMV seronegative adult volunteers. Vaccine2009;28:484–93

11. Precopio ML, Betts MR, Parrino J, Price DA, Gostick E, Ambrozak DR,et al. Immunization with vaccinia virus induces polyfunctional andphenotypically distinctive CD8(þ) T cell responses. J Exp Med 2007;204:1405–16

CMV DC Vaccines Enhance Polyfunctionality of ATCT in Glioma

www.aacrjournals.org Cancer Res; 78(1) January 1, 2018 263

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

12. Seder RA, Darrah PA, RoedererM. T-cell quality inmemory and protection:implications for vaccine design. Nat Rev Immunol 2008;8:247–58

13. Nebbia G, Mattes FM, Smith C, Hainsworth E, Kopycinski J, Burroughs A,et al. Polyfunctional cytomegalovirus-specific CD4þ and pp65 CD8þ Tcells protect against high-level replication after liver transplantation. Am JTransplant 2008;8:2590–9

14. Ding ZC, Huang L, Blazar BR, Yagita H, Mellor AL, Munn DH, et al.Polyfunctional CD4(þ) T cells are essential for eradicating advanced B-celllymphoma after chemotherapy. Blood 2012;120:2229–39

15. Mitchell DA, Batich KA, Gunn MD, Huang MN, Sanchez-Perez L, Nair SK,et al. Tetanus toxoid and CCL3 improve dendritic cell vaccines inmice andglioblastoma patients. Nature 2015;519:366–9

Cancer Res; 78(1) January 1, 2018 Cancer Research264

Reap et al.

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469

2018;78:256-264. Published OnlineFirst November 1, 2017.Cancer Res   Elizabeth A. Reap, Carter M. Suryadevara, Kristen A. Batich, et al.   T Cells That Target Cytomegalovirus in GlioblastomaDendritic Cells Enhance Polyfunctionality of Adoptively Transferred

  Updated version

  10.1158/0008-5472.CAN-17-0469doi:

Access the most recent version of this article at:

  Material

Supplementary

  http://cancerres.aacrjournals.org/content/suppl/2017/10/31/0008-5472.CAN-17-0469.DC1

Access the most recent supplemental material at:

  Cited articles

  http://cancerres.aacrjournals.org/content/78/1/256.full#ref-list-1

This article cites 15 articles, 5 of which you can access for free at:

  Citing articles

  http://cancerres.aacrjournals.org/content/78/1/256.full#related-urls

This article has been cited by 2 HighWire-hosted articles. Access the articles at:

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  [email protected]

To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/78/1/256To request permission to re-use all or part of this article, use this link

on October 28, 2020. © 2018 American Association for Cancer Research. cancerres.aacrjournals.org Downloaded from

Published OnlineFirst November 1, 2017; DOI: 10.1158/0008-5472.CAN-17-0469