644. Decreased Generation and Cytotoxic Function of Adenovirus-Specific CD8 T Cells in Both CD28 and FasL Deficient Mice

Abstract

The development of conditionally replicative adenoviruses (CRAds) has presented a new platform for cancer gene therapy. The relative contribution of CRAd-related toxicity, and the cytotoxic immune response, especially the antigen specific cytotoxic T lymphocyte (CTL) response, in tumor cell killing is unclear. The anti-CRAd CTL may destroy tumor cells by specific killing; on the other hand, the spread of viral progeny may be limited by the clearance of infected cells. A better understanding of the effector mechanism of the CTL response would have implications in design of adenovirus-derived vectors, in particular, the potent CRAds. Previous studies indicated that FasL and TNF-α but not perforin, play a central role in immune-mediated adenoviral vector clearance, whereas CD28 is thought to be a critical molecule for the generation of effector cells. However, it has been difficult to clearly distinguish effector cell generation from its function due to the inability to specifically recognize Ad-specific CTLs. To investigate the effector mechanisms directly, we employed two novel approaches: tetramer staining and an antigen specific in vivo CTL assay. The Db-E1Bp MHC-I tetramer, that recognizes the immunodominant epitope of Ad E1B (192 VNIRNCCYI) is the first Ad-mouse tetramer. B6 mice, CD28KO mice and FasL deficient gld mice which are all on the H-2b background, as well as D2 mice (H-2d), were immunized with 4×108 i.u. of wild type Ad5. Eight days later, target cells from syngeneic mice were labeled with high concentration of CFSE (2μM; CFSEHi) and pulsed with E1Bp, or labeled with low concentration of CFSE (0.2μM; CFSELo), but unpulsed. Immunized mice were then adoptively transferred with identical numbers of CFSEHi and CFSELo cells. After 6 hours, lymphocytes were isolated from the spleen and liver and stained with fluorescent anti-CD8 and anti-CD44 antibodies and the Db-E1Bp tetramer. Flow cytometry results showed that in B6 mice, approximately 33% of the CD8+ T cells in liver and 11% in spleen are Db-E1Bp positive. The E1Bp-specific killing in the spleen of B6 mice was 93% whereas no Db-E1Bp+ CD8+ T cells or E1Bp-specific killing were observed in D2 mice. E1Bp-specific lysis was significantly correlated with the frequency of Db-E1Bp+CD8+ T cells in B6 mice (p<0.01). As expected, CD28 deficient mice exhibited an 80% reduction in generation of Db-E1Bp+CD8+ efffector T cells and E1Bp specific killing efficiency was reduced by 65% compared to B6 mice. Surprisingly, the E1Bp-specific killing efficiency in gld mice was only minimally reduced by 25% compared to B6 mice. There was also a 25% decrease in generation of Db-E1Bp+CD8+ T cells in gld mice. However, the specific cytotoxicity of individual CTL in both CD28KO and gld mice was compromised in that the percentage lysis per cell was lower than in B6. Our results indicate 1) CD28 plays an important role in generation of the specific CTL response to Ad; 2) FasL-dependent pathway plays a minor role in CTL response and also affects generation of CTLs; and 3) tetramer analysis of Ad-specific CTL and specific in vivo CTL assay can be used to distinguish generation and effector functions of CTLs.