The mechanism whereby activated human T cells trigger inflammatory cytokine production by monocytes

It is now widely accepted that T lymphocytes can participate in the elimination of tumor cells. T cells were shown to mediate tumor rejection in mice, and the availability of long-term clonal cultures of murine anti-tumor cytolytic T lymphocytes (CTL) allowed to dissect their antigenic repertoire and eventually to identify their target antigens. In humans, CTL clones that specifically lyzed autologous tumor cells could be derived from tumor-infiltrating lymphocytes, which had shown to have anti-tumor activity upon adoptive transfer together with IL-2, or from blood lymphocytes. The identification of the different types of tumor-specific antigens that were recognized by such CTL clones, encoded either by genes that were expressed in tumor cells but silent in normal cells, by genes that were mutated in the tumor cells, by genes that were overexpressed in the tumor cells as compared to normal cells, or by melanocytic differentiation genes in the case of melanomas, demonstrated the capacity of T lymphocytes to kill tumor cells specifically. These results are at the root of current T cell directed cancer immunotherapies: immunization towards tumor-specific antigens, adoptive transfer of tumor-specific T cells, and antibody blockade of T cell inhibitory signals. When we analyzed the antitumor T cell responses of melanoma patients who displayed tumor regressions following vaccination with defined tumor-specific antigens, we did not find the expected strong anti-vaccine CTL responses. Instead, we found much stronger responses against tumor-specific antigens absent from the vaccine, and these T cells were also present in regressing metastases. Our observations suggested that few antivaccine T cells sparked the priming or restimulation of other antitumor T cells that constituted the bulk of the tumor rejecting effectors. We feel that this sparking process, resulting in antigen spreading and clonal spreading (164), might be important in all T cell directed cancer immunotherapies, including adoptive transfers which can show clinical efficacy even though the transferred CD4+ cells are not lytic, and in which both clonal spreading and antigen spreading take place. The present work is part of our efforts to understand the mechanisms of this sparking process. A possible explanation is that killed tumor cells release antigens that can then be cross-presented by dendritic cells. We believe that this cannot be the sole explanation, because the release of antigens by tumor cells is likely to happen in many other circumstances such as chemotherapy, or necrosis by hypoxia. We reasoned that the sparking process might include inflammatory mediators that would attract new immune cells into the tumor. Since we had observed that in melanoma metastases T cells are almost invariably accompanied by macrophages, the latter could be the source of these inflammatory mediators. Thus we set out to examine whether tumor-specific CTL could, upon killing melanoma cells, stimulate monocytic cells to release inflammatory mediators. Understanding the ‘spark’ will help to improve the immunization modalities that could induce such tumor microenvironment resulting in a higher proportion of clinically responding patients. I have shown here that activated T cells stimulate the production of proinflammatory cytokines by monocytes. The first part of the work focused on the identification of the T cell factors responsible for this stimulation of pro-inflammatory cytokine production, more precisely the factors that triggered the secretion of the two main pro-inflammatory cytokines IL-1β and IL-6. First I showed that both cell-contact and soluble factors produced by the activated T cells contributed to IL-1β and IL-6 induced-secretion. The soluble factors alone were able to trigger a production of low levels of IL-1β and IL-6, and cell-contact increased the secretion. The CD40/CD40L interaction appeared to be the main cell contact-dependent stimulus of IL-1β /IL-6 production, while TNFα, GM-CSF, and IFNγ were the main soluble factors that stimulated monocytic cells to secrete IL-1β and IL-6. While TNFa alone was a weak inducer of IL-1β secretion, the TNFα/GM-CSF combination increased IL-1β production. The addition of IFNγ to the TNFα/GM-CSF combination was required for IL-6 secretion by THP-1 cells. In the second part of the work I explored the mechanism behind the synergy between TNFα and GM-CSF to stimulate the production of IL-1β by THP-1. This synergy was observed at the transcriptional level. TNFα activates the NF-kB signaling pathway, which was a prerequisite for IL1B transcription. GM-CSF enhanced the TNFα effect on IL1B transcription through a JAK2-dependent process. Downstream of JAK2 the STAT5 and the MAPK pathway are activated. Although JAK2 and STAT3 are also activated in THP-1 after G-CSF treatment, this cytokine had no effect on the TNF-induced IL1B transcription. Therefore I explored the possibility that STAT5 was involved in the stimulation of IL1B transcription. I could not detect a binding of STAT5 to the IL1B promoter. Another possible explanation is an involvement of the ERK/MAPK pathway, which is activated by GM-CSF and TNFα Eeperiments with pharmacological inhibitors suggested the participation of the ERK/MAPK pathway, but not that of p38, JNK and PI3K, in the increased IL1B transcription stimulated by the TNFα/GM-CSF combination.(BIFA - Sciences biomédicales et pharmaceutiques) -- UCL, 201