Development of an Escherichia coli expressing listeriolysin-O vaccine against Wilms tumor gene 1-expressing tumors.

Through their ability to induce cytotoxic T-lymphocytes and inhibit Foxp3 T-regulatory cells, Escherichia coli expressing listeriolysin-O (LLO) and a model tumor antigen have been shown to exert strong antitumor activity. The aim of this study is to extend these observations to a self-protein and clinically relevant tumor antigen associated with most types of adult leukemia: Wilms tumor gene 1 (WT1). We demonstrate that an E. coli coexpressing LLO and WT1 is capable of inducing a strong antitumor effect against WT1-expressing tumors in vivo through its ability to induce cytotoxic T-lymphocytes and inhibit the function of Foxp3 T-regulatory cells. Furthermore, we have characterized the immunodominant epitope involved in this effect (NAPYLPSCL) and demonstrated that coinjection of NAPYLPSCL with E. coli-LLO resulted in an antitumor effect largely equivalent to that obtained with E. coli-LLO/WT1. Our data demonstrate that the results obtained with a clinically irrelevant model tumor antigen remain valid with a "real" tumor antigen and that the adjuvant properties of the E. coli-LLO vaccine can be exploited in conjunction with peptides. The results obtained in this study will facilitate the translation of this work to human studies by combining antigenic motifs relevant to specific human leukocyte antigen haplotypes with the adjuvant effect of E. coli-LLO.

Listeriolysin O expressed in a bacterial vaccine suppresses CD4+CD25high regulatory T cell function in vivo.

CD4(+)CD25(high) regulatory T cells (Treg) protect the host from autoimmune diseases but are also obstacles against cancer therapies. An ideal cancer vaccine would stimulate specific cytotoxic responses and reduce/suppress Treg function. In this study, we showed that Escherichia coli expressing listeriolysin O and OVA (E. coli LLO/OVA) demonstrated remarkable levels of protection against OVA-expressing tumor cells. By contrast, E. coli expressing OVA only (E. coli OVA) showed poor protection. High-avidity OVA-specific CTL were induced in E. coli LLO/OVA-vaccinated mice, and CD8(+) depletion--but not NK cell depletion, abolished the antitumor activity of the E. coli LLO/OVA vaccine. Phenotypic analysis of T cells following vaccination with either vaccine revealed preferential generation of CD44(high)CD62L(low) CD8(+) effector memory T cells over CD44(high)CD62L(high) central memory T cells. Unexpectedly, CD4(+) depletion turned E. coli OVA into a vaccine as effective as E. coli LLO/OVA suggesting that a subset of CD4(+) cells suppressed the CD8(+) T cell-mediated antitumor response. Further depletion experiments demonstrated that these suppressive cells consisted of CD4(+)CD25(high) regulatory T cells. We therefore assessed these vaccines for Treg function and found that although CD4(+)CD25(high) expansion and Foxp3 expression within this population was similar in all groups of mice, Treg cells from E. coli LLO/OVA-vaccinated animals were unable to suppress conventional T cells proliferation. These findings provide the first evidence that LLO expression affects Treg cell function and may have important implications for enhancing antitumor vaccination strategies in humans.