The immunology of DNA vaccines.

The surprising observation that direct inoculation of an expression plasmid encoding a foreign protein into the skin of mice resulted in the induction of antibody responses, demonstrated that injection of "naked" DNA could result in antigen expression in an immunogenic form (1). This observation and the subsequent demonstration that intramuscular injections of plasmid DNA encoding influenza nucleoprotein could protect mice against a challenge with live influenza virus have opened up new avenues for vaccine development (2-3). Immunization with plasmid DNA has been shown to activate both humoral and cellular immune responses, including the generation of antigen-specific CD8(+) cytotoxic T cells as well as CD4(+) T helper cells (4). An increasing number of studies using experimental animal models have demonstrated that plasmid DNA immunization can promote effective immune responses against numerous viruses, including influenza, rabies, HIV, HBV, HCV, and HSV; several bacteria, including: Mycobacterium tuberculosis, Mycoplasma pulmonis, and Borrelia burgdorferi; as well as parasites, such as malaria and leishmania (4). Phase I clinical vaccine trials are currently being performed for HIV, HBV, and influenza virus. With the molecular identification of tumor antigens (5), there has also been increasing interest in the development of DNA-based immunization for cancer. Preclinical studies demonstrate that DNA-based immunizations targeting model tumor antigens such as chicken ovalbumin (6), ß-galactosidase (7), or CEA (8) induce protective immune responses leading to rejection of a subsequent, normally lethal challenge with antigen-expressing tumor cells.

Partial purification of murine tumor-associated peptide epitopes common to histologically distinct tumors, melanoma and sarcoma, that are presented by H-2Kb molecules and recognized by CD8+ tumor-infiltrating lymphocytes.

The B16/BL6 melanoma is a relatively nonimmunogenic tumor expressing low levels of MHC class I molecules. BL6 clones expressing transfected H-2Kb class I molecules were, by contrast, highly immunogenic in immunocompetent mice. Tumor-infiltrating lymphocytes (TILs) generated from the H-2Kb+ BL6 lesions (Thy 1.2+, CD8+, CD4-) efficiently lysed H-2Kb+ melanoma (CL8-1 and 2Kb38) and the H-2Kb+ nonrelated 3-methylcholanthrene (MCA)-105 sarcoma, but not the H-2Kb- parental melanoma BL6-8. This strongly suggests that CL8-1, 2Kb38, and MCA-105 express identical or cross-reactive T cell epitopes recognized by CL8-1 TILs in the context of the H-2Kb class I allele. To identify the T cell epitopes, peptides were acid-eluted from various cells, and fractionated by HPLC. Five HPLC fractions (F1mel-F5mel) of 70 tested contained peptides derived from H-2Kb+ CL8-1 melanoma (but not H-2Kb- melanomas) that were capable of conferring susceptibility to CL8-1 TIL lysis on H-2b-expressing target cells (EL4, C1R.Kb), but not on H-2d-expressing P815 target cells. CL8-1 TILs failed to recognize peptides derived from H-2Kb+ nonmelanoma targets such as EL4 or normal B6 splenocytes. Interestingly, CL8-1 TILs appeared to recognize peptide species contained in two HPLC fractions derived from the MCA-105 sarcoma (F1sar and F5sar). Conversely, TILs derived from MCA-105 lesions recognized MCA-105 and CL8-1 tumor cells, as well as F5mel and F5sar peptides presented by EL4 targets. These data support common murine tumor-associated peptide epitopes presented by H-2Kb and recognized by CD8+ CTLs derived from histologically distinct tumors, melanoma and sarcoma.