Molecular Immunology Group

CD8+ Cytotoxic T lymphocytes (CTL) play crucial roles in the eradication of cells that are transformed (e.g. cancer) or infected with intracellular pathogens (e.g. viruses). Their coreceptor CD8, consisting of a disulfide linked α and β chain, plays crucial roles in the thymic selection of CD8+ T cells, their differentiation and activation of effector functions. Activation of these cells relies on engagement of their T cell antigen receptor (TCR) by cognate Major Histocompatibility Complexes (MHC) peptide complexes on antigen presenting (APC) or target cells. By binding to TCR associate MHC-peptide complexes CD8 strengthens TCR ligand binding and brings CD8-associated p56lck (lck) to TCR/CD3, thereby promoting phosphorylations of immunoreceptor tyrosine-based activation motifs (ITAM) and initiation of diverse down stream signaling cascades, e.g. involving the tyrosine kinase ZAP-70, the linker of activation of T cells (LAT), mobilization of intracellular calcium, etc. Our research interests are focused on a better understanding of T cell responses and how they can be generated and modulated for most effective eradication of tumors and viral infections. In order to find novel strategies, basic research is conducted to understand in more depth the molecular mechanism underlying antigen recognition, the roles of CD8, thymic selection and the formation of the TCR repertoires. Main lines of research are described below.

1. Impact of CD8β on CD8+ T cells functions

We have observed previously that CD8β KO mice clear acute viral infections such as influenza or lymphocytic choriomeningits virus (LCMV) as efficiently as wild type mice (C57BL/6). This was surprising in view of previous findings showing that CD8β endows CD8 with efficient co-receptor functions, including strengthened CD8 association with the Src tyrosine kinase lck, lck activation and lck-mediated phosphorylation of TCR/CD3 ITAM, subsequent recruitment and activation of ZAP-70, phosphorylation of PLCγ and mobilization of intracellular calcium. This signaling pathway is central to antigen recognition by CTL and is a pre-requisite for CTL degranulation, i.e. perforin/granzyme mediated, the central cytotoxic mechanism on which clearance of acute viral infections as well as tumor eradication relies. Consistent with previous studies on T cell hybridomas, we now observed that in CTL from CD8β- mice this signalling axis was severely compromised; notably the intracellular calcium mobilization was dramatically reduced both in magnitude and duration.

At variance with this we have previously shown that LCMV elicited CTL from CD8β KO mice are perfectly able to kill target cells as assessed in a 51Cr release assay (1). To investigate this paradox we used another cytotoxic assay, namely CD107a up-regulation. CD107a (LAMP-1) contained in CTL granules and upon degranulation becomes surface expressed, and therefore is a reliable measure of CTL degranulation and hence perforin/granzyme-mediated cytotoxicity. CTL from CD8β KO mice were negative in CD107a cytotoxic assays. Further investigations revealed that CD8+ T cell from CD8β KO mice expressed ~10-fold higher levels of FasL (CD95L), as assessed by flow cytometry and real time PCR. Not surprisingly then, the cytotoxicity mediated by CD8β- CTL turned out to be essentially mediated by Fas/FasL. Thus, whereas normal CD8β+ CTL can also kill via Fas/FasL, this cytotoxicity is typically slower, CD8 independent and secondary to granzyme/perforin-mediated cytotoxicity.

We therefore hypothesized that introduction of CD8β in CD8+ T cells that were selected in its absence might substantially increase their global effector functions. To examine this, we established an adoptive transfer system, in which CD8β- T cells from CD8β KO mice expressing as transgene (tg) the β chain of the Db/gp33-specific P14 TCR were transduced in vitro with CD8β, transferred into C57BL/6 (B6) mice and these then infected with LCMV. The responses of the transferred and endogenous CD8+ T cells were discerned by using different CD45 alleles (CD45.2 for transferred and CD45.1 for host cells). Transferred Db/gp33 tetramer+ CD8αβ+ T cells after LCMV infection (d 8) exhibited the same highly increased FasL expression as CD8β- cells, indicating that it not affected by restoring CD8β expression. Moreover, these CD8β transduced cells upon incubation with Db/gp33+ target cells displayed restored CD107a expression and calcium responses, i.e. normal perforin/granzyme-mediated killing. These results indicate that introduction of CD8β in CD8+ T cells selected in its absence provides a means to generate super CTL that in addition to normal CTL functions have strongly increased FasL- mediated cytotoxicity. In cancer as well as in persistent viral infections, CD8+ T cells fail to eradicate transformed, respectively infected cells and there is ample evidence for dysfunctions in perforin/granzyme-mediated cytotoxicity. We therefore investigated whether introduction of CD8β in T cells that have been selected may be a means to overcome these defects (see sections 4). It is noted that this strategy in addition to dramatically boosting FasL expressing on CD8+ T cells, and hence Fas/FasL-mediated killing also involves usage grossly different TCR repertoires (see section 2).

2. High resolution TCR repertoire analysis and what one can learn from it

CD8+ T cell responses arising in CD8β KO mice are CD8 independent, i.e. their functional responses are not affected by blocking anti-CD8 antibodies or CD8 binding ablating mutations of the target cells restricting MHC I molecule (1). We therefore reasoned that different, namely CD8 independent TCR were selected in the thymus of CD8β KO mice. By using anti-TCR Vβ (TRBV) -specific antibodies and flow cytometric analysis, we however observed only minor differences in the TRBV usage of LCMV immune Db/gp33-specific CD8+ T cells (1). Because CD8+ T cells are selected in the thymus and TCRα chain rearrangement occurs on CD4, CD8 double positive (DP) thymocytes concomitant to positive selection, we hypothesized that compromising the integrity of CD8 (by deletion of CD8β) is more likely to impact on the repertoire of the TCRα chain.

To conclusively investigate this, we generated mouse lines that expressed as transgene (tg) the TCRβ chain of the Db-restricted lymphocytic choriomeningitis virus ((LCMV) gp33-specific P14 TCR (P14β) and CD8β or not. Due to allelic exclusion in these mice the TCR repertoire is solely determined by the one of the TCRα chain. In the mouse the there are about 100 TCR Vα (TRAV) and 60 TCR Jα (TRAJ) genes, allowing a priori some 6000 possible TRAV-TRAJ recombinations. In order to conclusively analyse TCRα chain repertoires, we established a novel method that allows deep sequencing of TCRα chain transcripts (tens of millions of complete sequences) that were amplified in unbiased manner. Analysis of naïve and LCMV Db/gp33 immune (d 8) CD8+ T cells revealed that CD8β deficiency narrowed the TCRα chain repertoire dramatically on immune, less so on naïve CD8+ T cells. The TCRα chain repertoire of CD8β- T cells essentially comprised a fraction of the one CD8β+ T cells, with only few CD8β- unique sequences, many of which were CDR3α-specific (Fig.1). In the repertoires of Db/gp33-specific cells the TRAV14D-1_TRAJ48 recombination (IMGT nomenclature, see http://imgt.cines.fr/textes/IMGTrepertoire/LocusGenes/) appeared with high frequencies, namely on CD8β- cells, making up 74% of all TRAV_TRAJ recombinations. Remarkably, for this recombination frequency of bona fide P14 TCRα chain CDR3α sequence was high (88.2% on CD8β- cells and 43.6% on CD8β+ cells) (Fig. 1 and unpublished data).

Furthermore, because this novel TCR repertoire analysis provides unprecedented high resolution we will also use it to investigate in more depth the mechanism of TCRα gene rearrangements about much has been speculated based on limited data available. One popular view is the coordinate TCR gene rearrangement model according to which recombinations of 3’ proximal TRAV with 5’ proximal TRAJ near the T early activation (TEA) promoter occur with higher incidence (Fig. 2). Our data so far are at variance with this model, indicating instead that there is no difference in frequencies TRAV-TRAJ recombinations involving TRAV from the original 3’ or duplicate 5’ locus and that the TRAJ usage correlates with the TRAV usage, irrespective of their genomic location.

This strategy is equally applicable to the analysis of TCRβ chain repertoires and to species other than mice. Making use of this, in one project we investigate the inter-relationship and “sharing of labor” between the TCRα and β repertoires in shaping the pre-immune as well as immune TCR repertoires. In another project we will use integrated, longitudinal TCRα and β repertoire analysis on CD4+ and CD8+ T cells from cancer patients before and after treatments, such as CTLA-4 blockade and/or caner vaccination.

3. Impact of CD8β on the thymic selection of CD8+ T cells

It is widely accepted that the pre-immune TCR repertoire is essentially generated in the course of thymic selection. By using TCR transgenic, RAG d efficient mice and panels of peptide variants spanning a range of binding affinities, an universal affinity threshold has been indentified that defines the windows of positive and negative selection (2). As it has been shown that CD8αβ, but not CD8αα substantially strengthens TCR-MHC-peptide binding, we addressed the question what impact CD8β on the thymic selection in S14 TCR tg, RAG KO mice. The S14 TCR is Kd-restricted and specific for PbCS(ABA) (SYIPSAEK(ABA)I (ABA: 4-azido benzoic acid)-specific). Previous studies on thymic selection in this (and other) system have shown that there is an affinity threshold (KD ~ 6 μM) for thymic selection. As we have shown previously that in the absence of CD8β the affinity of MHC-peptide for S14 TCR dramatically decreases, we now studied the thymic selection using foetal thymic organ cultures (FTOC) and a range of PbCS(ABA) peptide variants. We observed a shift in the hierarchy of selecting ligand, i.e. that the wt agonist (4P, i.e SYIPSAEK(ABA)I) was selecting in CD8β KO FTOC, whereas in the wt FTOC the next weaker peptide variant (4L, i.e. (SYILSAEK(ABA)I) was. Related shifts were observed when CD69 up-regulation was assessed, which is an indicator for negative thymic selection. Results so far indicate that in the absence of CD8β there is a modest shift in the hierarchy of selecting ligands towards higher affinity. However, we suspect that other factors are involved as well, such as e.g. up-regulation of FasL and possibly FasL signalling to compensate for the blunted lck , ZAP-70 signalling pathway inferred to by deletion of CD8β.

4. New roads in generating tumorcidal T cell responses

An attractive strategy to eradicate or to contain tumor consists in generating tumor-specific T cell-responses, namely tumor-associated antigen (TAA)-specific effector CD8+ T cells. While such approaches unlike chemo- or irradiation therapy are minimally invasive, they are prone to diverse tumor escape mechanisms; by which tumors evade immunity by exploiting mechanisms like peripheral tolerance, cell or cytokine mediated suppression or blunting effector functions that normally serve to prevent auto-immunity. We investigate different strategies to subvert tumor escape.

First, we hypothesize that introduction of CD8β in CD8+ T cell that have been selected in its absence, substantially increases their effector functions and that this may be exploited to eradicate cancer, by subverting tolerance to cancer. To test this, we established an adoptive transfer system in which cells from P14β tg, CD8β KO mice were transduced in vitro with CD8β and upon transfer in B6 mice challenged by LCMV infection. As control CD8αβ+ T cells from P14β tg mice were included. The responses of the transferred and endogenous CD8+ T cells were discerned by the use of different CD45 alleles (CD45.2 for transferred and CD45.1 for host cells). The transferred Db/gp33 tetramer+ CD8αβ+ T cells after LCMV infection still exhibited greatly increased FasL expression, demonstrating that it is not reversed by restored CD8β expression and hence is a consequence of thymic selection. At the same time these CD8β transduced cells exhibited restored perforin/granzyme responses, indicating that introduction of CD8β in CD8+ T cells selected in its absence provides a means to generate super CTL that in addition to normal CTL functions have strongly increased FasL- mediated cytotoxicity. Moreover, we argue that because in CD8β mice CD8+ T cells selected that are CD8 independent and have a markedly different TCR repertoire than wild type mice, restored CD8β expression is likely to further heighten their reactivity. Wile this harbours the risk to elicit auto-immune reactions, it seems highly likely to subvert tumor-related tolerance. We now are testing the capacity of such adoptively transferred cells to eradicate established tumors that express as surrogate TAA gp33, including the highly malignant mouse melanoma B16. Diversifications of this strategy include the use of: i) CD8β fused with yellow fluorescent protein (YFP), ii) genetically engineered CD8β with enhanced co-receptor functions and iii) normal CD8β KO mice, i.e. without the P14 TCRβ tg, to generalize the finding obtained.

Second, we established a pre-clinical mouse tumor model to establish optimal TAA-specific cell responses and to study their interaction with established TAA+ tumors in vivo. To this end we used HLA-A*0201, HLA-DRB1*0101 tg, H-2-/- (A2/DR1) mice and the cancer testis (CT) TAA NY-ESO-1 (ESO), which is highly immunogenic, broadly expressed on a range of tumors and for which numerous immonogenic epitopes have been mapped (3). In order to identify an optimal vaccine regime, we first immunized mice with different A2, and DR1-restricted ESO peptides in CpG and IFA. The ESO peptides 127-137 and 157-165 were thus found to elicit the strongest DR1 and A2-restricted CD4 and A2 T cell responses, respectively. Using these peptides a highly efficient immunizations scheme was established in which mice were first immunized to generate a strong ESO 123-137-specific Th1 response; subsequently these mice were inoculated with mature dendritic cells pulsed with this and the ESO157-165 peptide, resulting in efficient priming of ESO-specific CD8+ T cells in lymph nodes and egress of highly active ESO-specific CTL in the periphery (3). Mechanistic studies revealed that the β chemokines CLL3/CCL3 and their common receptor CCR5 were instrumental in orchestrating cellular interactions and the Th1 cytokines IL-2 + IFNγ in CD8+ T cells priming and sphingosine-1-phosphate (S1P1) for their entry into circulation (3).

We have derived from A2/DR1 mice a fibrosarcoma, and from this cells lines that were transfected or not with ESO. This system is not used investigate the key factors and variables that govern tumor eradication of established ESO+ tumors in vivo. In addition to direct therapeutic vaccination we also use adoptive transfer of in vivo primed CD8+ T cells in tumor bearing hosts, which allows their prior in vitro manipulations, e.g. to allow in vivo tracking of transferred cells or re-program their trafficking, namely promoting tumor targeting.

Selected Publications

1. Angelov, G.S., Guillaume, P., Cebecauer M., Bosshard, G., Dojcinovic, D., Baumgaertner, P, and Luescher, I.F. 2006. Soluble MHC-peptide complexes containing long rigid linkers abolish CTL-mediated cytotoxicity. J. Immunol. 176:3356-65.

2. Naeher D, Daniels MA, Hausmann B, Guillaume P, Luescher I, Palmer E. 2007. A constant affinity threshold for T cell tolerance. J. Exp. Med. 204:2553-2559.

3. Johannsen A, Legler DF, Luther SA, Luescher IF. 2010. Definitions of key variables for the induction of optimal NY-ESO-1-specific CD8+ T cell responses in HLA transgenic mice. J Immunol. 185:3445-3455.

4. Angelov G. S., Guillaume P, and Luescher, I.F. 2009. CD8 knockout mice mount normal anti-viral CD8+ T cell responses – but why ? Int. Immunol. ;21:123-135.

5. Guillaume P, Dojcinovic D, Luescher IF. 2009. Soluble MHC-peptide complexes: tools for the monitoring of T cell responses in clinical trials and basic research. Cancer Immun. 25;9:7.

6. Guillaume P, Baumgaertner P, Neff L, Rufer N, Wettstein P, Speiser DE, Luescher IF. 2010. Novel soluble HLA-A2/MELAN-A complexes selectively stain a differentiation defective subpopulation of CD8+ T cells in melanoma patients. Int J Cancer. 127:910-23.

7. Ayyoub M, Dojcinovic D, Pignon P, Raimbaud I, Schmidt J, Luescher I, Valmori D. 2010. Monitoring of NY-ESO-1 specific CD4+ T cells using molecularly defined MHC class II/His-tag-peptide tetramers. Proc Natl Acad Sci U S A. 107:7437-42.

8. Ayyoub M, Dojcinovic, D, Pignon, P, Raimbaud, I, Old LJ, Luescher I, Valmori D. 2010. Monitoring of tumor-specific CD4+ T cells using molecularly defined MHC class II/His tag-peptide tetramers. Clin. Cancer Res. 16:4607-4615.

Danijel Dojcinovic, Research assistant

Raphaël Genolet, Postdoctoral Fellow

Philippe Guillaume, Assistant Investigator

Luca Cariolato, Postdoctoral Fellow

Julie Leignadier, Postdoctoral Fellow

Emily Navid, Technician

Julien Schmidt, Postdoctoral Fellow