T lymphocytes and skin cancer
T-cells are white blood cells representing an essential part of the immune system. They recognize and destroy any diseased cell which has been altered, e.g. by infection with a virus or by transformation into cancer cells. Melanoma has proven to be one of the most immunogenic among human tumors. A relatively high number of molecularly defined antigens that are recognized by T lymphocytes have been identified in the last 15 years. For instance, the differentiation antigen Melan-A is selectively expressed by normal cells of the melanocytic lineage including skin melanocytes and by a large majority of malignant melanomas. In the human melanoma cancer model, tumor antigen-specific T-cell responses often develop, but their protective effects fail to control or eliminate the disease. The main goal of the projects described below is to improve the possibilities for using tumor-specific cytolytic T-cells in anti-cancer therapy.
T lymphocytes bear antigen-specific T-cell receptors
Each T-cell possesses a unique entity called T-cell receptor. The main function of this receptor is to recognize target molecules of viral- or tumor-origin. While well-defined antiviral-specific T-cell responses consist of T-cells expressing T-cell receptors endowed with high avidity for their antigen ligand, T-cells directed against tumor antigens express T-cell receptors of lower avidity. The relative lack of high avidity T-cell receptors may be a major reason why immune responses towards tumor progression remain inefficient.
A major avenue for cancer immunotherapy is the adoptive transfer of in vitro expanded tumor reactive T-cells. Although many aspects of T-cell fate and persistence following transfer of these cells to cancer patients are poorly understood, it is expected that transfer of T-cells expressing high avidity T-cell receptors would improve therapeutic benefit. Alternatively, since the antigen specificity of T-cells is determined solely by the T-cell receptor molecules, it has become increasingly feasible to transfer high avidity T-cell receptors to polyclonal unselected T-cells of the patient. Nevertheless, major areas still remain to be investigated to greatly facilitate the design of a set of "off the shelf" tumor reactive T-cell receptors for gene transfer and subsequent adoptive transfer. These include (1) identifying the type of T-cells that are most efficient in anti-cancer therapy, (2) isolating the T-cell clones with high avidity T-cell receptors for tumor antigen and, (3) understanding the structural basis of T-cell receptor avidity for tumor antigen.
Tracking tumor-specific T lymphocytes in melanoma patients
To specifically address these points, we developed a novel approach that allow us to carefully dissect and follow human immune responses against tumors (Rufer, 2005; Rufer et al., 2005; Rufer et al., 2003). We are now able to isolate anti-tumor T-cells bearing T-cell receptors with high potency for tumor recognition and destruction. Using this new strategy, we have recently tracked and followed dominant tumor-specific T cell clones in sevela melanoma patients over extended period of time (Appay et al., 2006; Derre et al., 2007; Derre et al., 2008; Speiser et al., 2006; Voelter et al., 2008a; Voelter et al., 2008b). Overall, our molecular-based strategy allows for the first time to precisely assess T-cell responses ex vivo by dissecting distinct subsets and anatomical compartments, over time, and during subsequent immunotherapy. The identification of those "super" anti-tumor selected T-cell receptors and their subsequent fine characterization is a key step to better understand the functional requirements of T cell receptors to efficiently recognize tumor cells.
Defining tumor-reactive T cells in patients after therapeutic vaccination
Therapeutic vaccines aim at inducing and boosting specific T cells mediated immunity to reduce tumor burden. Such approaches provide a unique opportunity to study T-cell priming and memory formation in humans. CpG are synthetic oligodeoxynucleotides (ODN) that contain unmethylated CpG motifs similar to those observed in bacterial DNA. CpG ODN elicit a complex immuno-modulatory cascade that includes the production of pro-inflammatory cytokines, and the stimulation of dendritic cell activation through TLR9 triggering. We recently conducted a phase I clinical trial in which melanoma patients were vaccinated with low dose of CpG mixed with tumor antigen and Incomplete Freund's adjuvant. Rapid and strong tumor-reactive T-cell responses were observed in 27/28 patients demonstrating that CpG ODN is an efficient vaccine adjuvant, to the extent that the applied vaccine formulation clearly elicits higher T-cell frequencies than any other T-cell vaccine currently applied in humans. In fact, this is the first time that vaccine induced T-cell responses reach T-cell frequencies and T-cell functions at levels comparable to those of virus specific T-cells that successfully protect from viral diseases (Speiser et al., 2008; Speiser et al., 2005).
Model for the generation of a protective immune response
Previous reports revealed large TCR repertoires among T-cells responding to a given tumor antigen. Here, we were able to identify dominant T-cell clonotypes because we focused our analysis on distinct T-cell subsets (i.e. intermediate versus differentiated cells). In particular, we observed a highly restricted TCR repertoire in the most differentiated T-cell compartment, supporting the notion that restricted but dominant antigen-specific T-cell responses are composed of functionally differentiated effector cells with efficient cytolytic properties. Thus, high-resolution characterization of dominant tumor-reactive T-cell responses provides the basis to identify biological parameters associated with protective T-cell immunity. Moreover, the mechanisms involved in the selection of particular T-cell clonotypes and the impact of vaccination can now be precisely assessed. In line with these observations, we propose the following working model in which protective immune responses against tumor antigens may involve the selection and generation of differentiated T-cells with potent effector functions, restricted clonal diversity and increased avidity of their T-cell receptors (fig. 1). This type of immune response is typically observed following chronic viral infections with i.e. Epstein-Barr virus (EBV) or Cytomegalovirus (CMV). If our model were to be confirmed, the development of new vaccines allowing the generation of such efficient tumor-specific immune T-cell responses will be of central importance for cancer therapy. One of our objectives is to understand the mechanisms by which immune responses to EBV and CMV can provide life long protection in healthy carriers, while such responses against melanoma are induced and improved by vaccination, but remain largely inefficient in eliminating tumors.
Working model for the generation of a protective immune response upon CD8+ T-cell differentiation. This involves the selection and generation of dominant differentiated cytolytic T lymphocytes with potent effector functions, restricted clonal diversity and increased avidity of their TCRs. CMV; cytomegalovirus.
Towards improvement of structural and functional basis of TCR avidity for tumor antigen
In contrast to viral-specific T-cell bearing TCRs with high avidity for their ligand, CD8+ T cells directed against tumor antigens express TCR of lower avidity. It is therefore tempting to reengineer the TCR sequence to increase its affinity for common tumor epitopes. To address this point, we are currently using a rational in silico approach in which the TCR sequence is optimized step by step in a predictive and thus very controlled manner. As proof-of concept, we have chosen to study the NY-ESO-1/A2 system, because of its clinical relevance (immunodominant T-cell clone found in a long survivor patient; LAU 155) and for the availability of a very closely related X-ray structure. This provides, therefore, an ideal system to study the effect of CDR point mutations on cytolytic T lymphocytes via lentiviral transduction. Optimized TCR candidates are further tested in redirected TCR gene transfer experiments on bulk CD8+ T lymphocytes for TCR cell surface expression, TCR-pMHC off rates using multimer decay assays, proliferation and killing assays. Altogether, combined bio-modeling and functional experiments allows for the first time to specifically assess the impact of single amino acid replacements within the TCR CDR2 region on both TCR-pMHC affinity and functional avidity. Here we show that the controlled introduction in the TCR sequence of only one or two amino acid substitutions can have drastic positive or negative repercussions on the avidity and functionality of the cells. Such studies open the way towards the generation of T cell receptors with improved avidity for tumor antigens, thus facilitating the design of tumor-reactive TCRs for gene transfer and subsequent adoptive transfer for the treatment of melanoma.