Broad Spectrum Tumor Antigen-Specific Cytotoxic T Lymphocytes (CTL) for Therapy of Hematological Malignancy.

Abstract 4083

Poster Board III-1018

Tumor immunotherapy using cytotoxic T lymphocytes (CTLs) can effectively treat EBV-associated tumors such as Hodgkin's Lymphoma (HL). To extend the benefit of CTL therapy to other non-viral associated tumors we have now developed a means of generating CTLs that recognize a wide spectrum of non-viral tumor-associated antigens (TAA) expressed on a range of hematological malignancies, including Survivin, MAGEA4, Synovial sarcoma X (SSX2), WT1, Prame, Proteinase 3 (PR3), as well as TAAs such as Alpha-fetoprotein (AFP), MAGEA1 and MAGEA3, which are expressed on solid tumors.

The generation of TAA-CTL is complicated by the low frequency of circulating reactive T cells which are often anergized or tolerized. We therefore assessed the potency of different cytokine combinations to overcome T cell anergy. The cytokines tested were Th1 polarizing (IL12, IL18, IL27), pro-survival/-proliferative (IL7, IL15, IL4), and Treg inhibitory (IL-6). We first used autologous dendritic cells (DCs) pulsed with overlapping peptide libraries (pepmixes) spanning the different antigens to stimulate T cells in the presence of the optimal cytokine combination. To further optimize the potency of this approach we chose to use combinations of antigens to produce multi TAA-CTL. The use of pepmixes spanning entire antigens as a stimulus both increases the range of suitable patient HLA polymorphisms beyond those matched to individual peptide epitopes and to reduce the risk of tumor immune evasion that is often observed when single defined-epitope TAA peptides are used as a stimulus. Antigen combinations were chosen based on their tumor expression profile; e.g. to target lymphoma we grouped SSX2, Survivin, and MAGEA4, for leukemia we targeted WT1, Prame, PR3, and Survivin and for hepatocellular carcinoma we combined MAGEA1, MAGEA3 and AFP.

Using lymphoma as a model system we were consistently able to generate TAA CTL directed against SSX2, Survivin, and MAGEA4 (n=6 healthy donors). The lines were polyclonal comprising CD4+ (mean 57+/-7%) and CD8+ (mean 40+/-16%) T cells, and showed reactivity against SSX2 (median 345 spot forming units (SFU)/1×10⁶ CTL, range 20-4065), Survivin (median 290 SFU/1×10⁶ CTL, range 30-4632.5) and MAGEA4 (median 255 SFU/1×10⁶ CTL, range 27.5-2962.5), using the relevant pepmixes as a stimulus in ELIspot assays. The CTL were also cytolytic against autologous pepmix-pulsed targets; SSX2 (mean 53+/- 32% specific lysis) and Survivin (30%) at an effector:target (E:T) ratio of 40:1, as well as against whole antigen-expressing targets (30% killing of fibroblasts expressing Survivin from a retrovirus vector, E:T 100:1) with no recognition of non-transduced control targets (9%). For leukemia we generated CTL lines with reactivity against WT1 (mean 733 SFU/1×10⁶ CTL +/-88), Prame (mean 555 SFU/1×10⁶ CTL +/-326), PR3 (mean 145 SFU/1×10⁶ CTL +/-156), and Survivin (mean 32 SFU/1×10⁶ CTL +/-25), n=2. We were similarly successful in generating multi TAA-CTL targeting hepatocellular carcinoma with reactivity for MAGEA1, MAGEA3 and AFP (mean 740+/-969, 929+/-624, 163+/-117 SFU/1×10⁶ CTL, respectively), (n=2).

These multi TAA-CTL were generated using pepmixes, which may not be well suited to clinical use. To more accurately target tumors expressing naturally processed antigens, we reproduced these experiments using DCs nucleofected with tumor antigen-encoding DNA plasmids. Using this approach we were consistently able to generate polyclonal SSX2- and MAGEA4-specific T cells. Production of Survivin-reactive cells required a plasmid encoding a Survivin-ubiquitin fusion protein to enhance proteosomal processing, degradation and subsequent presentation. By significantly increasing expression of CD70 on DCs [from median 23.4% (range 13.3-48.8) to median 61.9% (range 46.7-83.8)], (p=0.025) using nucleofection (n=4), we could also enhance the production of in vitro expanded MAGEA3-specific T-cells (from 822 SFU/1×10⁶ CTL to 2460 SFU/1×10⁶ CTL). Finally, we successfully substituted CD40L activated autologous B-blasts for DCs to reduce the volume of patient blood required and to facilitate the logistics of preparation.

This study shows the successful generation of tumor specific CTL lines targeting multiple tumor antigens simultaneously to reduce the risk of tumor immune escape and describes a technology feasible for clinical application targeting a broad spectrum of hematologic malignancies.