Adenoviral gene transfer into dendritic cells efficiently amplifies the immune response to LMP2A antigen: A potential treatment strategy for Epstein-Barr virus-positive Hodgkin\'s lymphoma

Int. J. Cancer: 93, 706 –713 (2001) © 2001 Wiley-Liss, Inc.

Publication of the International Union Against Cancer

ADENOVIRAL GENE TRANSFER INTO DENDRITIC CELLS EFFICIENTLY AMPLIFIES THE IMMUNE RESPONSE TO LMP2A ANTIGEN: A POTENTIAL TREATMENT STRATEGY FOR EPSTEIN-BARR VIRUS–POSITIVE HODGKIN’S LYMPHOMA Benedikt GAHN1, Fernando SILLER-LOPEZ1, Angela D. PIROOZ1, Eric YVON1, Stephen GOTTSCHALK1, Richard LONGNECKER2, Malcolm K. BRENNER1, Helen E. HESLOP1, Estuardo AGUILAR-CORDOVA1 and Cliona M. ROONEY1* 1 Center for Cell and Gene Therapy, Baylor College of Medicine, Houston, TX, USA 2 Robert H. Lurie Cancer Center, Northwestern University, Chicago, IL, USA The EBV– encoded LMP2A protein is consistently expressed in EBVⴙ Hodgkin’s lymphoma and can be targeted by CTLs. CTLs stimulated conventionally by LCLs have little activity against LMP2Aⴙ target cells. Here, we describe an alternative approach, based on the in vitro stimulation of CTLs with DCs genetically modified with 2 E1/E3-deleted recombinant adenoviruses, AdGFPLMP2A, encoding a fusion gene of GFP and LMP2A, and AdLMP2A, encoding LMP2A only. Transduction of DCs with AdGFPLMP2A at MOI 1,000 resulted in LMP2A expression in up to 88% of DCs. LMP2A protein was expressed in 40% of DCs transduced with AdLMP2A at an MOI of 100. Higher MOI resulted in DC death. CTL lines activated by transduced DCs had a higher frequency of LMP2A tetramer-specific CTLs than CTL lines activated by LCLs. CTLs stimulated with transduced DCs lysed both autologous fibroblasts infected with vaccinia virus LMP2A (FBvaccLMP2A) and autologous LCLs, which express LMP2A at lower levels. In contrast, CTLs generated from the same donors by stimulation with autologous LCLs showed minimal lysis of FBvaccLMP2A. Moreover, 1 donor who did not respond to LMP2A when CTLs were stimulated with LCLs became a responder when LMP2A was expressed by transduced DCs. Hence, recombinant adenoviruses encoding LMP2A effectively transduce DCs and direct the generation of LMP2A-specific CTLs. This approach will be a potent strategy in Hodgkin’s lymphoma immunotherapy. © 2001 Wiley-Liss, Inc. Key words: Hodgkin’s lymphoma; Epstein-Barr virus; LMP2A; dendritic cells; recombinant adenovirus; cytotoxic T lymphocytes

Improvements in radiochemotherapy have correspondingly improved the prognosis of patients with Hodgkin⬘s lymphoma.1 However, patients resistant to the standard therapeutic approaches have a poor outcome. Moreover, both life expectancy and quality of life of patients cured of Hodgkin’s lymphoma are significantly reduced by treatment-related mortality and morbidity.2,3 These limitations of current treatment protocols illustrate the need for more effective and less toxic therapeutic approaches. In Hodgkin⬘s lymphomas, up to 49% of specimens have been shown to carry EBV DNA and to express EBV genes.4,5 Adoptive immunotherapy with donor-derived, EBV-specific CTLs has significantly improved the clinical outcome of patients with EBV⫹ immunoblastic lymphoma after T cell– depleted allogeneic stemcell transplantation.6,7 In an ongoing study, our group has successfully generated EBV-specific CTLs in patients with EBV⫹ Hodgkin’s lymphoma.8 After infusion, these CTLs home to the tumor sites, persist in the circulation for up to 9 months and produce transient clinical benefits. In this protocol, LCLs are used as EBV antigen-presenting cells. LCLs activate polyclonal CTL populations that are preferentially directed against the immunodominant EBNA3A, -3B and -3C EBV proteins.9 These immunogenic proteins are not expressed in HRS cells. Instead, the EBV antigens on HRS cells are restricted to the expression of a subset of latent proteins, EBNA1, LMP1, LMP2A and BARF0.10 –12 LCLs have limited efficacy at stimulating CTLs directed against these subdominant proteins.

Targeting the EBV proteins expressed in EBV⫹ HRS cells is further complicated by their limited immunogenicity. EBNA1 is not processed for HLA class I presentation due to an internal glycine–alanine repeat region.13 LMP1-specific CTL clones are rare in EBV⫹ donors,14 and very few LMP1 epitopes have been identified.15 Moreover, LMP1 displays heterogeneity between viral strains,16 and CTLs raised against B cell (B95-8 prototype) derived LMP1 may not recognize LMP1 tumor variants.17 BARF0 is a protein expressed from the BamHI A rightward transcript. Attempts to identify BARF0-specific CTLs have produced lymphocytes that recognized target cells expressing BARF0 at high levels but failed to recognize the antigen expressed at lower levels in LCLs.18 LMP2A epitopes were shown to be conserved among Hodgkin’s lymphoma biopsy samples displaying little heterogeneity between viral strains.19,20 Also, most donors have a low but measurable frequency of circulating LMP2A-specific CTLs that can be activated and expanded in vitro.14 Hence, LMP2A may be the protein of choice to be targeted by CTLs in patients with Hodgkin’s lymphoma. To stimulate LMP2A-specific CTLs, LMP2A peptide-pulsed DCs have been successfully used, but this approach is limited to known MHC-restricted peptides.21,22 A more promising strategy is the production of CTLs by genetically modified DCs that direct the CTL response to virally transduced genes.23–25 This approach allows expression of the whole protein, leading to presentation of multiple, undefined antigen epitopes.26 CTLs specific for LMP2B have been amplified using DCs transduced with a recombinant adenovirus encoding for LMP2B.27 Although LMP2B shares 8 exons with LMP2A, the first exon of LMP2A encodes a unique 119 amino acid tail that is absent from LMP2B.28 Therefore, LMP2A will contain additional CTL epitopes.29 To date, however, it has not proved possible to generate adenoviral vectors expressing full-length LMP2A. We have used homologous recombination in Escherichia coli to prepare an E1/E3-deleted recombinant adAbbreviations: CTL, cytotoxic T lymphocyte; DC, dendritic cell; EBV, Epstein-Barr virus; HRS cells, Hodgkin-Reed-Sternberg cells; iu, infective units; LCL, lymphoblastoid cell line; MAb, monoclonal antibody; MACS, magnetic cell sorting; MCM, monocyte conditioned medium; MOI, multiplicity of infection; PBMC, peripheral blood mononuclear cell; pfu, plaque-forming units; PHA, phytohemagglutinin. Grant sponsor: Deutsche Krebshilfe; Grant sponsor: National Institutes of Health; Grant numbers: ROI CA 74126; CA 61384; Grant sponsor: Baylor College of Medicine. *Correspondence to: Center for Cell and Gene Therapy, Baylor College of Medicine, 6621 Fannin Street, Houston, TX 77030, USA. Fax: ⫹832-825-4732. E-mail: [email protected] Received 26 January 2001; Revised 6 April 2001; Accepted 10 April 2001 Published online 13 June 2001

GENERATION OF LMP2A-SPECIFIC CTL

enoviral vector with a GFPLMP2A fusion gene (AdGFPLMP2A). In addition, for future clinical trials in Hodgkin’s lymphoma patients, a recombinant adenovirus encoding LMP2A alone (AdLMP2A) was constructed. When used for transduction of DCs, both vectors were capable of inducing CTLs that lysed LMP2Aexpressing target cells more effectively than CTLs stimulated with LCLs. MATERIAL AND METHODS

Cell lines and blood donors The human embryonic kidney cell line 293 was purchased from the ATCC (Bethesda, MD). The cell line was maintained in GVL-modified MEM/MAb medium (Hyclone, Logan, UT). All other lines were derived from 3 healthy EBV⫹ donors, A, B and C, of known HLA type (A, HLA-A2,31,B57,62; B, HLA-A2,B27,51; C, HLA-A24,B48,61). B-LCLs were established, using a standard protocol,8 by infecting PBMCs with the B95-8 supernatant in cyclosporin A (Novartis, East Hanover, NJ). Autologous PHAactivated blasts were prepared by stimulating PBMCs with PHA 5 ␮g/ml (Sigma, St. Louis, MO) and cultured for at least 7 days in RPMI-1640 (GIBCO BRL, Grand Island, NY), 10% FCS (Hyclone) and 2 mM L-glutamine (GIBCO BRL) (complete medium). Primary autologous skin fibroblasts were isolated by punch biopsies taken from the volar surface of the forearm. Tissue was incubated in 6-well plates in complete medium with fungizone 1 ␮g/ml (GIBCO BRL). Once a primary culture was established, cells were subsequently maintained in complete medium and passaged at a 1:3 dilution when confluent. Construction of the recombinant adenoviruses AdGFPLMP2A and AdLMP2A To generate the GFPLMP2A fusion cDNA (2,632 bp), a 1.8 kb fragment containing the LMP2A cDNA was removed from pBluescriptLMP2A and cloned into pEGFP-C1 (Clontech, Palo Alto, CA). The conjugated product was inserted into the shuttle vector pShCMV30 to yield the plasmid pShGFPLMP2A. pShLMP2A was constructed in an identical fashion by ligating LMP2A cDNA into pShCMV. Homologous recombination between pShGFPLMP2A or pShLMP2A and pAdEasy-1 resulted in the recombinant adenoviral plasmids pAdGFPLMP2A and pAdLMP2A. 293 cells were transfected with pAdGFPLMP2A or pAdLMP2A, and individual plaques were isolated, amplified, purified by CsCl density gradient centrifugation and subsequently dialyzed. Titers of large-scale stocks were assigned on 293 cells after vector exposure by GFP expression and plaque formation, respectively. Functional titers of recombinant viruses were 4 ⫻ 109 iu/ml (AdGFPLMP2A) and 6 ⫻ 109 pfu/ml (AdLMP2A), equivalent to 3 ⫻ 1011 viral particles/ml and 4 ⫻ 1011 viral particles/ml, respectively. Recombinant viruses were stored at – 80°C until use. Generation of DCs DCs were generated as described by Butterfield et al.,23 with some modification. Briefly, PBMCs were purified by Ficoll (Lymphoprep; Nycomed, Oslo, Norway) gradient separation. Mononuclear cells (4 ⫻ 107) were plated in RPMI-1640 with 1% human AB serum (BioWhittaker, Walkersville, MD) (RPMI-AB) in a T-25 flask (Costar, Corning, NY) for 2 hr at 37°C in a humidified CO2 incubator. Non-adherent cells were removed by rinsing with PBS, and loosely adherent cells were cultured in RPMI-AB with 800 U/ml GM-CSF (Sargramostim Leukine; Immunex, Seattle, WA) and 500 U/ml IL-4 (Genzyme, Cambridge, MA) for 7 days. IL-4 and GM-CSF were added on days 2, 4 and 6. On day 7, cells were harvested by vigorous washing. After transduction, cells were matured by adding 25% allogeneic MCM for another 48 hr and then used for CTL stimulation. MCM was prepared as described previously,31 with minor modifications. Briefly, 1 ⫻ 107 PBMCs were put into an Ig-coated bacteriologic plate (Falcon, Becton Dickinson, Franklin Lakes, NY) for 1 hr at 37°C in 8 ml of RPMI-1640 with 10% human AB serum (RPMI-10%AB). Nonadherent cells were rinsed away and adherent cells cultured in

707

fresh RPMI-10%AB at 37°C. After 24 hr, conditioned medium was removed and sterile filtered. Adenoviral transduction Immature DCs were harvested on day 7 with cold PBS (GIBCO BRL), resuspended in 500 ␮l RPMI-1640, added to 24-well plates and transduced with AdGFPLMP2A at a multiplicity of infection (MOI) of 1,000 iu/1 ⫻ 106 cells or with AdLMP2A at an MOI of 100 pfu/1 ⫻ 106 cells, unless otherwise stated. Plates were spun for 90 min at 1,000 g.32 Afterward, DCs were cultured in 25% MCM supplemented with 500 U/ml IL-4 and 800 U/ml GM-CSF and then used to stimulate CTLs. RNA isolation and RT-PCR RNA was extracted from AdGFPLMP2A-infected DCs and uninfected control DCs (2 ⫻ 106) 48 hr after transduction using the Trizol method (GIBCO BRL, Gaithersburg, MD) according to the manufacturer’s instructions. First-strand cDNA was synthesized using the superscript pre-amplification system (GIBCO BRL) with the supplied oligo(dT)12–18 primers. One microliter of the reverse-transcription reaction product was used for subsequent PCR. The sequence of the human LMP2A primers, which generated a 260 bp fragment, was as follows: sense 5⬘-CAG TTA GTA CCG TTG TGA CC-3⬘, anti-sense 5⬘-CTA GTA TCA GGA GCA CAA GC-3⬘. The sequence of the GAPDH primers was as follows: sense 5⬘-GCA TTG CTG ATG ATC TTG AGG CT-3⬘, anti-sense 5⬘CAC CCA TGG CAA ATT CCA TGG C-3⬘. Samples were denatured for 40 sec at 94°C, annealed for 2 min at 64°C and extended for 2 min at 72°C, for a total of 35 cycles. The PCR product was subjected to electrophoresis on a 1.5% agarose gel containing ethidium bromide and visualized by UV illumination. GAPDH was used as an internal control for RNA integrity. Primers were synthesized by Life Technologies (Gaithersburg, MD). Flow cytometry Expression of the LMP2A protein was analyzed by FACS on day 9 (2 days after transduction). The anti-LMP2A MAb, derived from clone 8c3, was kindly provided by E. Kremmer (Munich, Germany). Briefly, transduced and control DCs were fixed for 10 min on ice with 4% paraformaldehyde, permeabilized for 30 min with 1% saponin (Sigma) and stained with primary rat antiLMP2A MAb.33 Phycoerythrin (PE)– conjugated goat anti-rat IgG1 MAb (Caltag, San Francisco, CA) was used as secondary antibody. Non-specific binding was measured using rat isotype control (Caltag) as primary antibody, followed by staining with secondary PE-conjugated anti-rat IgG. Samples were acquired on a FACScan flow cytometer (Becton Dickinson, Mountain View, CA) and the data analyzed using CellQuest software (Becton Dickinson). GFP transgene expression was assessed by flow cytometry with the instrument setting for fluorescein detection. Expression of the surface molecules was measured on non-fixed, non-permeabilized DCs using PE-conjugated MAbs: anti-CD3, -CD16, -CD19, -CD56, -CD1a, -CD80, -CD83, -CD86 and -DR PerCP (Becton Dickinson). FITC-, PE- and PerCP-conjugated, isotype-matched mouse IgGs were used as controls (Pharmingen, San Diego, CA). CTL lines were analyzed with anti-CD8 FITC, -CD56 PE, -CD3 PerCP, -CD4 PE, -CD16 FITC and -CD19 PE (Becton Dickinson). T cells CD8⫹ cells were selected from non-adherent cells of donors A and B using MACS depletion of CD4⫹, CD14⫹, CD19⫹ and CD56⫹ cells according to the manufacturer’s instructions (Miltenyi Biotech, Bergisch Gladbach, Germany). Stimulation of CTLs with DCAdGFPLMP2A or DCAdLMP2A CD8⫹ cells (1 ⫻ 106) were co-cultured with 5 ⫻ 104 autologous, transduced or non-transduced DCs in 2 ml of complete medium in the presence of 10 ng/ml IL-7 (R&D Systems, Minneapolis, MN). Cultures were restimulated on day 10 and after that weekly with irradiated autologous DCAdGFPLMP2A or

708

GAHN ET AL.

DCAdLMP2A at a responder-to-stimulator ratio of 20:1. IL-2 (20 U/ml, Proleukin; Chiron, Emeryville, CA) was started 3 days later and added twice weekly. Stimulation of CTLs with LCLs PBMCs (2 ⫻ 106/well) were co-cultured with 5 ⫻ 104 autologous LCLs in 2 ml of complete medium. Cultures were restimulated on day 10 and after that weekly with irradiated autologous LCLs at a responder-to-stimulator ratio of 4:1. IL-2 (20 U/ml, Proleukin; Chiron) was started 3 days later and added twice weekly. Tetramer staining Soluble HLA-A0201-CLGGLLTMV PE-conjugated tetramers were prepared by the National Institute of Allergy and Infectious Diseases (NIAID) tetramer core facility (Atlanta, GA). The CLGGLLTMV peptide was synthesized by M. Campbell (Houston, TX). CTLs (1 ⫻ 106) were incubated on ice for 30 min in PBS with 1% FCS containing the PE-labeled tetrameric complex. Samples were additionally incubated with anti-CD8 FITC and antiCD3 PerCP. IgG–PE was used for isotype control. Stained cells were fixed in PBS containing 0.5% paraformaldehyde. Samples were analyzed by FACS as described. Preparation of fibroblast targets for cytotoxicity assay Autologous fibroblasts were infected for 1 hr with 5 pfu/cell of recombinant vaccinia virus containing LMP2A (FBvaccLMP2A) or the control gene LacZ (vaccLacZ) encoding for the ␤-galactosidase (all gifts from E. Kieff, Boston, MA).9 Fibroblasts (2 ⫻ 106) were infected with each construct and incubated with 0.1 mCi 51 Cr. Autologous fibroblasts constitutively expressing GFP were generated by transducing cells with a retroviral vector encoding GFP (RetroGFP). Transduced cells were separated by FACS and further expanded. Autologous fibroblasts were also transduced with an adenoviral vector encoding ␤-galactosidase (AdLacZ). All fibroblasts were treated for 24 hr with 100 U/ml IFN-␥ (Pharmingen) before use in the cytotoxicity assay. Cytotoxicity assay CTLs were tested for specific cytotoxicity against autologous fibroblasts, either uninfected or infected with vaccLMP2A, vaccLacZ, RetroGFP or AdLacZ. Autologous LCLs, HLA class I–mismatched LCLs and autologous PHA blasts were also tested. Target cells (1 ⫻ 106) were labeled with 0.1 mCi 51Cr and mixed with various numbers of effector cells to give E:T ratios of 40:1, 20:1, 10:1 and 5:1. Anti-MHC class I blocking antibody (Dako, Glostrup, Denmark) was added to control wells to determine whether recognition of target cells by CTLs was MHC class I–restricted. Target cells incubated in complete medium or 5% Triton X-100 (Sigma) were used to determine spontaneous and maximal 51Cr release, respectively. After 4 (LCLs, PHA blasts) or 5 (fibroblasts) hr, supernatants were collected and radioactivity was measured on a gamma counter. The mean percentage of specific lysis of triplicate wells was calculated as 100 ⫻ (experimental release – spontaneous release)/(maximal release – spontaneous release). RESULTS

Maturation of DCs in MCM After 7 days of culture of adherent PBMCs in GM-CSF and IL-4, 50% of cells were large and granular (Fig. 1a). These large, granular cells, gated in Fig. 1a, were 90% lineage-negative, DRpositive and 98% negative for CD83, representing an immature DC phenotype (Fig. 1b). These cells were transduced with AdGFPLMP2A and AdLMP2A and cultured for 2 more days in MCM. Percentages of large, granular and of lineage-negative, DR-positive cells in the resulting population were unchanged. However, up to 88% of large, granular cells expressed CD83 (Fig. 1c), indicating maturation of DCs. There was also an increased level of DR expression in the CD83⫹ population, further confirm-

ing maturation of DCs. These data show that maturation of DCs is not inhibited by adenoviral transduction. LMP2A mRNA and protein expression in AdGFPLMP2A- and AdLMP2A-transduced DCs To determine the transcription of LMP2A mRNA, RT-PCR was performed on RNA from AdGFPLMP2A-transduced DCs (data not shown). DCs were infected with AdGFPLMP2A at MOI 100 and left until 48 hr post-infection, at which time RNA was extracted. In untransduced DCs, LMP2A mRNA was not detected. In DCs transduced with AdGFPLMP2A, LMP2A mRNA was present but not when RT-PCR was performed on transduced DCs without reverse transcriptase. A GAPDH fragment served as internal control. These data indicate the transcription of LMP2A mRNA from the adenovirus vector. To determine the level of protein expression, DCs were infected with AdGFPLMP2A at MOI 100, 300, 500 and 1,000. The cells were harvested 48 hr post-infection and analyzed for LMP2A expression. In FACS analysis, levels of adenovirus-derived LMP2A protein increased in a dose-dependent manner across the range of MOIs tested. LMP2A was expressed in 39% of DCs at MOI 300, 63% at MOI 500 and 88% at MOI 1,000 (Fig. 2). FACS analysis for GFP performed in parallel showed similar expression levels for the GFP protein, as expected. There was no significant toxicity at any of these MOI values, as determined by cell counts and trypan blue staining. DCs were also transduced with AdLMP2A. At MOI 100, 40% of AdLMP2A-transduced DCs expressed LMP2A protein. Using this vector at higher MOI did not result in higher transduction rates but led to increased DC death: 90% of DCs died at MOI 300. AdGFPLMP2A- and AdLMP2A-transduced DCs amplify LMP2A-specific CTLs To generate a specific CTL response against LMP2A, CD8⫹ lymphocytes from donor A were stimulated with AdGFPLMP2Atransduced autologous DCs. In donor B, CD8⫹ cells were stimulated with AdLMP2A-transduced DCs. In donor A, FACS analysis of the resulting T-cell line showed 99% positivity for CD8, 35% positivity for both CD8 and CD56 and 1% positivity for CD4 cells. In donor B, 98% of cells stained positive for CD8, 7% for CD8 and CD56 and 0.4% for CD4. To estimate the frequency of LMP2A-specific CTLs after stimulation with transduced DCs, we used CLGGLLTMV-A2 tetramers representing 1 LMP2A epitope. In unstimulated peripheral blood, 0.01% (Fig. 3a) and 0.07% (Fig. 3d), respectively, of CD8⫹ T cells of donors A and B stained positive for the CLGGLLTMV-A2 tetramer. In donor A after stimulation with DCAdGFPLMP2A and in donor B after stimulation with DCAdLMP2A, 20% (Fig. 3b) and 11% (Fig. 3e), respectively, of CD8 cells reacted with the CLGGLLTMV-A2 tetramers. In both donors, stimulation of CD8⫹ cells with untransduced DCs did not expand CLGGLLTMV-A2 tetramer-specific CTLs (data not shown). When using autologous LCLs as stimulators, in donor A 0.01% (Fig. 3c) and in donor B 1.6% (Fig. 3f) of the resulting CTLs were specific for the LMP2A-A2 tetramer. The cytolytic activity of responder cells was tested against a panel of 51Cr-labeled autologous and allogeneic targets. In CTLs from donor A, 42% specific lysis of autologous fibroblasts infected with vaccLMP2A was observed at an E:T ratio of 40:1 (Fig. 4a). CTLs also lysed autologous LCLs. FBvaccLMP2A expressed LMP2A with high intensity, whereas LCLs expressed LMP2A at lower levels (data not shown).9 Similarly, CTLs from donor B lysed both autologous LCLs and vaccLMP2A-infected autologous fibroblasts (Fig. 4b). In both donors, 51Cr release from autologous LMP2A-expressing fibroblasts and LCLs was inhibited by antiHLA class I MAbs. To exclude the possibility that the 51Cr release may be due to vaccinia virus– directed CTLs, fibroblasts infected with recombinant vaccinia viruses encoding ␤-galactosidase were used as targets and were not recognized. Neither autologous PHA

GENERATION OF LMP2A-SPECIFIC CTL

709

FIGURE 1 – Phenotype of DCs before and after transduction and the addition of MCM. After 7 days of culture of adherent cells in IL-4 and GM-CSF, cells were labeled with anti-HLA-DR PerCP and anti-CD83 PE. By scatter profile, 50% of adherent cells were large and granular (a). Cells gated in A were 98% negative for CD83, representing immature DCs (b). These cells were transduced and incubated for 2 more days in MCM; 88% of the resulting large, granular cells expressed CD83, indicating DC maturation (c).

We also checked for the induction of GFP-specific CTLs by AdGFPLMP2A-transduced DCs. In donor A, we observed 25% lysis of fibroblasts transduced with RetroGFP at an E:T ratio of 40:1, indicating the presence of GFP-specific CTLs (Fig. 4c). There was little lysis of fibroblasts transduced with AdLacZ as a target for adenovirus-directed CTL: 11% in donor A (Fig. 4c) and 3% in donor B (Fig. 4d). We compared the effectiveness of AdLMP2A-transduced DCs with LCLs as activators of LMP2A-specific CTLs. In donor B, there was no lysis of FBvaccLMP2A by LCL-activated CTLs. However, these CTLs effectively lysed target cells infected with the immunodominant EBV antigens EBNA3B and EBNA3C (Fig. 5a). In contrast, CTLs stimulated with AdLMP2A-transduced DCs specifically lysed FBvaccLMP2A, while EBNA3B- and EBNA3C-expressing target cells were not recognized (Fig. 5b). FIGURE 2 – GFP and LMP2A protein expression in AdGFPLMP2Atransduced DCs investigated by FACS at different MOI values. Expression of both proteins increased in a dose-dependant fashion, depending on the MOI of the AdGFPLMP2A used for DC transduction. The intensity of LMP2A expression correlated (close to a 1:1 ratio) with GFP expression.

blasts nor untransduced fibroblasts, representing targets for autoreactivity, were killed. 51Cr release in HLA-mismatched LCLs was not observed. T-cell lines co-cultured with untransduced DCs expanded poorly and did not kill FBvaccLMP2A (data not shown).

Activation of LMP2A-specific CTLs by DCAdLMP2A from unselected PBMCs and expansion with LCLs When using CTLs in clinical protocols, it is desirable to minimize ex vivo manipulations. We therefore examined our ability to generate LMP2A-specific CTLs from unmanipulated PBMCs, thus avoiding depletion of CD4⫹, CD14⫹, CD56⫹ and CD19⫹ cells. We determined in addition whether after 2 specific stimulations of PBMCs with DCAdLMP2A, autologous LCLs could be used for subsequent CTL expansion. This was performed in 2 HLA-A2⫹ donors (A and B) and in 1 HLA-A24⫹ donor (C). FACS analysis of the resulting effector cells showed 95 ⫾ 4% CD8⫹ cells

710

GAHN ET AL.

FIGURE 3 – Percentages of LMP2A CLGGLLTMV-A2 tetramer-specific CTLs in donors A and B in starting PBMC populations (a,d), CTL populations after stimulation with transduced DCs (b,e) and CTL populations after stimulation with LCLs (c,f). In donor A, DCs were transduced with AdGFPLMP2A; in donor B, DCs were transduced with AdLMP2A.

(mean ⫾ SD), 3 ⫾ 4% CD4⫹ cells and 1 ⫾ 1% CD3–, CD56⫹ cells. Co-incubation with FBvaccLMP2A resulted in 52 ⫾ 17% specific lysis at a 40:1 E:T ratio. This 51Cr release was completely inhibited by HLA class I antibodies. Uninfected fibroblasts and fibroblasts infected with vaccLacZ were not lysed (Fig. 6). DISCUSSION

Potentially the most promising effector cells for adoptive immunotherapy in EBV⫹ Hodgkin’s lymphoma are CTLs directed

against the LMP2A protein, rather than LMP1, EBNA1 or BARF0. CTLs specific for BARF0 were unable to recognize cells expressing the antigen at low levels,18 EBNA1 is not processed for HLA class I presentation13 and LMP1 is frequently mutated.16 Although LMP2B-specific CTLs have been reported, this protein lacks a significant portion of the LMP2A protein that likely contains important CTL epitopes.28 CTLs specific for the LMP2A antigen were amplified with only limited success from patients with Hodgkin’s lymphoma using LCLs as stimulators,8 while the use of

GENERATION OF LMP2A-SPECIFIC CTL

711

FIGURE 4 – CTLs stimulated by DCAdGFPLMP2A (donor A) and DCAdLMP2A (donor B) specifically lysed both FBvaccLMP2A and autologous LCLs (a,b). Specific lysis was also observed when CTLs stimulated by AdGFPLMP2A-transduced DCs were incubated with FBRetroGFP, indicating the presence of GFP-directed CTLs (c). There was only low 51Cr release using FBAdLacZ as a target for adenospecific CTLs (c,d).

FIGURE 5 – EBV antigen specificity of donor B CTL lines. CTLs stimulated with LCLs were directed against EBNA3B and EBNA3C (a), whereas CTLs stimulated with DCAdLMP2A were specific for LMP2A (b).

712

GAHN ET AL.

FIGURE 6 – CTLs initially stimulated with DCAdLMP2A and subsequently expanded with LCLs specifically lysed FBvaccLMP2A. Mean ⫾ SD of experiments on 3 donors is shown.

DCs pulsed with peptides to stimulate LMP2A-specific CTLs has been limited to known MHC-restricted peptides.21,22 In the present study, we describe the construction of recombinant adenoviruses encoding the entire LMP2A protein, facilitating effective stimulation of LMP2A-specific CTLs with transduced DCs. Use of AdGFPLMP2A-transduced DCs was complicated by the generation of GFP-specific CTLs. Although it remains to be demonstrated how this GFP-expressing system performs in vivo, the immunogenicity of the GFP protein could limit the use of AdGFPLMP2A to pre-clinical applications.34 GFP-specific CTLs may also overwhelm LMP2A-specific CTLs, though this appears unlikely in EBV-seropositive donors. For these reasons, we constructed the untagged AdLMP2A. Although MOI ⬎100 pfu/ml resulted in high DC death, even at a transduction rate of 40% the resulting AdLMP2A-transduced DCs were equally capable of inducing an LMP2A-specific CTL population. Use of a complete protein recombinant adenovirus is advantageous over single peptides in that the host alleles are able to select the relevant HLA-restricted epitopes for presentation. In a previous study, gp100-specific tumor-infiltrating lymphocytes were able to recognize 5 different epitopes on the surface of cells transduced with a gp100-encoding adenovirus.35 In the present study, staining with a single HLA-A2 tetramer likely under-estimated the frequency of LMP2A-specific CTLs. Only 11% and 20%, respectively, of CTLs were specific for the CLGGLLTMV-A2 epitope. The remaining CTLs, not stained by the CLGGLLTMV-A2 tetramer, are likely specific for other, as yet undefined, LMP2A epitopes. In donor A, FACS analysis showed 2 separate CLGGLLTMV-A2 tetramer-positive populations, indicating different affinities of LMP2A-specific CTLs to the tetramer. Because no EBV⫹ Hodgkin’s lymphoma cell line exists, we tested the cytotoxic specificity of CTLs stimulated by AdGFPLMP2A- or AdLMP2A-transduced DCs against fibroblasts infected with vaccLMP2A. Fibroblasts expressing LMP2A were

specifically lysed. Since FBvaccLMP2A expresses LMP2A at high levels, we were also interested in the lysis of LCLs that express LMP2A with lower intensity. Autologous LCLs were also lysed, showing that the generated CTL populations were capable of lysing target cell populations that expressed LMP2A at lower (probably in tumor cells biologically more relevant) levels. CTLs stimulated by DCs transduced with AdGFPLMP2A and AdLMP2A were much more efficient at lysing LMP2A-expressing target cells than CTLs stimulated with autologous LCLs. CTLs stimulated with LCLs were predominantly directed against the immunodominant EBNA3 antigens and had little activity against LMP2-expressing target cells. LCLs in donor A did not stimulate CLGGLLTMV-A2-specific CTLs, and LCL-stimulated CTLs did not kill LMP2A-expressing fibroblasts. In contrast, transduced DCs activated LMP2A-specific CTLs, implying that transduced DCs overcome the non-responsiveness to LMP2A expressed by LCLs. We could not detect adenovirus-specific killing. This finding is in accordance with the results of other investigators, who did not find lysis of adenoviral targets27,36,37 or lysis only at low levels23,38 when testing CTLs stimulated with recombinant adenovirus-transduced DCs. Production of adenovirus-specific CTLs may require a different DC-stimulation protocol.39 Adenovirus-specific CTLs could be re-activated when DCs were co-cultured with CTLs directly after transduction. In our protocol, DCs were cultured after transduction for 2 days in MCM before co-culture with PBMCs, which allowed for maximal transgene expression and maturation but probably resulted in loss of adenoviral epitopes. Methods for LMP2A-specific CTL generation for clinical use should be simple, feasible and reproducible. Depletion of CD4⫹, CD14⫹, CD19⫹ and CD56⫹ cells requires clinical grade MAbs that are not available. Multiple stimulations of CTLs with DCs require large amounts of blood, likely not available in patients with relapsed disease. In preparation for clinical studies, we stimulated unselected PBMCs with DCAdLMP2A and subsequently expanded the resulting CTL lines with autologous LCLs. This proved to be very effective, the resulting effector cell population being 95% positive for CD8 and specifically lysing LMP2A-expressing target cells. Hence, these modifications efficiently avoid the depletion of CD4⫹, CD14⫹, CD56⫹ and CD19⫹ cells and reduce the volume of patient blood necessary to generate DCs. In conclusion, use of DCs transduced with LMP2A-encoding adenoviruses is an efficient means of stimulating LMP2A-specific CTLs and is thus useful for adoptive immunotherapy protocols in patients with EBV⫹ Hodgkin’s lymphoma. ACKNOWLEDGEMENTS

BG was supported by a grant from the Deutsche Krebshilfe. This work was also supported in part by National Institutes of Health (NIH) grants ROI CA 74126 and CA 61384 (CR), a Doris Duke Distinguished Clinical Scientist award (HH) and the Department of Pediatrics, Baylor College of Medicine (Houston, TX).

REFERENCES

1. 2. 3.

4. 5.

Engert A, Wolf J, Diehl V. Treatment of advanced Hodgkin’s lymphoma: standard and experimental approaches. Semin Hematol 1999; 36:282–9. Aisenberg AC. Problems in Hodgkin’s disease management. Blood 1999;93:761–79. Milligan DW, Ruiz DEM, Kolb HJ, Goldstone AH, Meloni G, Rohatiner AZ, et al. Secondary leukaemia and myelodysplasia after autografting for lymphoma: results from the EBMT. EBMT Lymphoma and Late Effects Working Parties. European Group for Blood and Marrow Transplantation. Br J Haematol 1999;106: 1020 – 6. Herbst H, Niedobitek G, Kneba M, Hummel M, Finn T, Anagnostopoulos I, et al. High incidence of Epstein-Barr virus genomes in Hodgkin’s disease. Am J Pathol 1990;137:13– 8. Pallesen G, Hamilton-Dutoit SJ, Rowe M, Young LS. Expression of

6. 7.

8.

9.

Epstein-Barr virus latent gene products in tumour cells of Hodgkin’s disease. Lancet 1991;337:320 –2. Rooney CM, Smith CA, Ng C, Loftin SK, Li C, Krance RA, et al. Use of gene-modified virus-specific T lymphocytes to control Epstein-Barr virus-related lymphoproliferation. Lancet 1995;345:9 –13. Heslop HE, Ng CYC, Li C, Smith CA, Loftin SK, Krance RA, et al. Long-term restoration of immunity against Epstein-Barr virus infection by adoptive transfer of gene-modified virus-specific T lymphocytes. Nat Med 1996;2:551–5. Roskrow MA, Suzuki N, Gan Y-J, Sixbey JW, Ng CYC, Kimbrough S, et al. EBV-specific cytotoxic T lymphocytes for the treatment of patients with EBV positive relapsed Hodgkin’s disease. Blood 1998; 91:2925–34. Murray RJ, Kurilla MG, Brooks JM, Thomas WA, Rowe M, Kieff E, et al. Identification of target antigens for the human cytotoxic T cell

GENERATION OF LMP2A-SPECIFIC CTL

10.

11. 12.

13.

14.

15.

16.

17.

18.

19.

20. 21. 22. 23.

24.

response to Epstein-Barr virus (EBV): implications for the immune control of EBV⫹ malignancies. J Exp Med 1992;176:157– 68. Herbst H, Dallenback F, Hummel M, Niedobitek G, Pileri S, MullerLantzsch N, et al. Epstein-Barr virus latent membrane protein expression in Hodgkin and Reed-Sternberg cells. Proc Natl Acad Sci USA 1991;88:4766 –70. Rickinson AB, Moss DJ. Human cytotoxic T lymphocyte responses to Epstein-Barr virus infection. Annu Rev Immunol 1997;15:405–31. Oudejans JJ, van den Brule AJ, Jiwa NM, de Bruin PC, Ossenkoppele GJ, van der Valk P, et al. BHRF1, the Epstein-Barr virus (EBV) homologue of the BCL-2 protooncogene, is transcribed in EBVassociated B-cell lymphomas and in reactive lymphocytes. Blood 1995;86:1893–902. Levitskaya J, Coram M, Levitsky V, Imreh S, Steigerwald-Mullen PM, Klein G, et al. Inhibition of antigen processing by the internal repeat region of the Epstein-Barr virus nuclear antigen-1. Nature 1995;375:685– 8. Sing AP, Ambinder RF, Hong DJ, Jensen M, Batten W, Petersdorf E, et al. Isolation of Epstein-Barr virus (EBV)-specific cytotoxic T Lymphocytes that lyse Reed-Sternberg cells: implications for immune-medicated therapy of EBV Hodgkin’s disease. Blood 1997;89: 1978 – 86. Khanna R, Burrows SR, Nicholls J, Poulsen LM. Identification of cytotoxic T cell epitopes within Epstein-Barr virus (EBV) oncogene latent membrane protein 1 (LMP1): evidence for HLA A2 supertyperestricted immune recognition of EBV-infected cells by LMP1-specific cytotoxic T lymphocytes. Eur J Immunol 1998;28:451– 8. Khanim F, Yao QY, Niedobitek G, Sihota S, Rickinson AB, Young LS. Analysis of Epstein-Barr virus gene polymorphisms in normal donors and in virus-associated tumors from different geographic locations. Blood 1996;88:3491–501. Trivedi P, Hu LF, Christensson B, Masucci MG, Klein G, Winberg G. Epstein-Barr virus (EBV)– encoded membrane protein LMP1 from a nasopharyngeal carcinoma is non-immunogenic in a murine model system, in contrast to a B cell– derived homologue. Eur J Cancer 1994;30A:84 – 8. Kienzle N, Sculley TB, Poulson L, Buck M, Cross S, Raab-Traub N, et al. Identification of a cytotoxic T-lymphocyte response to the novel BARF0 protein of Epstein-Barr virus: a critical role for antigen expression. J Virol 1998;72:6614 –20. Murray PG, Constandinou CM, Crocker J, Young LS, Ambinder RF. Analysis of major histocompatibility complex class I, TAP expression, and LMP2 epitope sequence in Epstein-Barr virus-positive Hodgkin’s disease. Blood 1998;92:2477– 83. Busson P, Edwards RH, Tursz T, Raab-Traub N. Sequence polymorphism in the Epstein-Barr virus latent membrane protein (LMP)-2 gene. J Gen Virol 1995;76:139 – 45. Redchenko IV, Rickinson AB. Accessing Epstein-Barr virus-specific T-cell memory with peptide-loaded dendritic cells. J Virol 1999;73: 334 – 42. Takahashi K, Honeyman MC, Harrison LC. Dendritic cells generated from human blood in granulocyte macrophage-colony stimulating factor and interleukin-7. Hum Immunol 1997;55:103–16. Butterfield LH, Jilani SM, Chakraborty NG, Bui LA, Ribas A, Dissette VB, et al. Generation of melanoma-specific cytotoxic T lymphocytes by dendritic cells transduced with a MART-1 adenovirus. J Immunol 1998;161:5607–13. Linette GP, Shankara S, Longerich S, Yang S, Doll R, Nicolette C, et al. In vitro priming with adenovirus/gp100 antigen-transduced dendritic cells reveals the epitope specificity of HLA-A*0201–restricted

⫹

25.

26.

27.

28. 29.

30. 31.

32. 33.

34. 35.

36.

37.

38.

39.

713

CD8 T cells in patients with melanoma. J Immunol 2000;164:3402– 12. Diao J, Smythe JA, Smyth C, Rowe PB, Alexander IE. Human PBMC-derived dendritic cells transduced with an adenovirus vector induce cytotoxic T-lymphocyte responses against a vector-encoded antigen in vitro. Gene Ther 1999;6:845–53. Heemskerk MH, Hooijberg E, Ruizendaal JJ, van der Weide MM, Kueter E, Bakker AQ, et al. Enrichment of an antigen-specific T cell response by retrovirally transduced human dendritic cells. Cell Immunol 1999;195:10 –7. Ranieri E, Herr W, Gambotto A, Olson W, Rowe D, Robbins PD, et al. Dendritic cells transduced with an adenovirus vector encoding Epstein-Barr virus latent membrane protein 2B: a new modality for vaccination. J Virol 1999;73:10416 –25. Sample J, Liebowitz D, Kieff E. Two related Epstein-Barr virus membrane proteins are encoded by separate genes. J Virol 1989;63: 933–7. Niedobitek G, Kremmer E, Herbst H, Whitehead L, Dawson CW, Niedobitek E, et al. Immunohistochemical detection of the EpsteinBarr virus-encoded latent membrane protein 2A in Hodgkin’s disease and infectious mononucleosis. Blood 1997;90:1664 –72. He TC, Zhou S, da Costa LT, Yu J, Kinzler KW, Vogelstein B. A simplified system for generating recombinant adenoviruses. Proc Natl Acad Sci USA 1998;95:2509 –14. Romani N, Reider D, Heuer M, Ebner S, Kampgen E, Eibl B, et al. Generation of mature dendritic cells from human blood. An improved method with special regard to clinical applicability. J Immunol Methods 1996;196:137–51. Nyberg-Hoffman C, Shabram P, Li W, Giroux D, Aguilar-Cordova E. Sensitivity and reproducibility in adenoviral infectious titer determination. Nat Med 1997;3:808 –11. Subklewe M, Chahroudi A, Schmaljohn A, Kurilla MG, Bhardwaj N, Steinman RM. Induction of Epstein-Barr virus-specific cytotoxic Tlymphocyte responses using dendritic cells pulsed with EBNA-3A peptides or UV-inactivated, recombinant EBNA-3A vaccinia virus. Blood 1999;94:1372– 81. Stripecke R, Carmen VM, Skelton D, Satake N, Halene S, Kohn D. Immune response to green fluorescent protein: implications for gene therapy. Gene Ther 1999;6:1305–12. Zhai Y, Yang JC, Kawakami Y, Spiess P, Wadsworth SC, Cardoza LM, et al. Antigen-specific tumor vaccines. Development and characterization of recombinant adenoviruses encoding MART1 or gp100 for cancer therapy. J Immunol 1996;156:700 –10. Morgan SM, Wilinson GW, Floettmann E, Blake N, Rickinson AB. A recombinant adenovirus expressing an Epstein-Barr virus (EBV) target antigen can selectively reactivate rare components of EBV cytotoxic T-lymphocyte memory in vitro. J Virol 1996;70:2394 – 402. Reed DS, Romero P, Rimoldi D, Cerottini JC, Schaack J, Jongeneel CV. Construction and characterization of a recombinant adenovirus directing expression of the MAGE-1 tumor-specific antigen. Int J Cancer 1997;72:1045–55. Philip R, Alters SE, Brunette E, Ashton J, Gadea J, Yau J, et al. Dendritic cells loaded with MART-1 peptide or infected with adenoviral construct are functionally equivalent in the induction of tumorspecific cytotoxic T lymphocyte responses in patients with melanoma. J Immunother 2000;23:168 –76. Smith CA, Woodruff LS, Kitchingman GR, Rooney CM. Adenoviruspulsed dendritic cells stimulate human virus-specific T-cell responses in vitro. J Virol 1996;70:6733– 40.