Blazing a trail in T-cell recognition

Most αβ⁺ T lymphocytes recognize peptide antigens presented by class I or class II molecules of the MHC. Hanada et al describe a novel αβ TCR whose ligand contains no MHC molecule, but rather comprises soluble TRAIL (TNF-related apoptosis-inducing ligand) bound to Death Receptor 4 (DR4).¹ 

T lymphocytes bearing heterodimeric αβ TCRs, which are generated during thymic development by rearrangement of noncontiguous gene segments encoded in the T-cell receptor α and β loci, comprise the majority of T cells in blood and lymph nodes. Positive and negative selection of thymocytes bearing αβ TCRs produces a naive αβ T-cell repertoire that collectively expresses 2-3 × 10⁶ distinct antigen receptors.²  The natural ligands for most αβ TCRs (including those that participate in thymic selection) are short peptides bound to the antigen groove of the highly polymorphic class I and class II molecules of the MHC. In addition, small numbers of T cells expressing αβ TCRs whose natural ligands are lipid molecules bound in the analogous antigen groove of monomorphic MHC class Ib molecules such as CD1 can be found in all individuals.

Rare exceptions to the rule that αβ T cells recognize small peptide or lipid moieties bound to the antigen groove of classic or nonclassical MHC molecules have been described. CD8⁺ αβ T-cell lines showing MHC-unrestricted recognition of a tandemly repeated peptide epitope in the extracellular domain of the mucin molecule MUC1, a type I transmembrane protein, have been generated in vitro,³  and direct recognition of Hfe, a MHC class Ib molecule that does not have any antigen-presenting function, and its human orthologue HFE by murine CD8⁺ αβ T cells has also been described.⁴  In this issue of Blood, Hanada et al¹  further expand the range of potential ligands for αβ TCRs. Through sequential expression screening of cDNA libraries, they demonstrate that the αβ TCR of a human CD4⁺ T-cell clone recognizes a molecular complex that contains neither a small peptide, nor lipid moiety, nor protein product of the MHC, but rather soluble TRAIL bound to DR4 (see figure).

Hanada et al had previously shown that this T-cell clone, isolated from an individual with renal cell carcinoma (RCC), displayed MHC-unrestricted recognition of a determinant expressed on most RCC lines, but not on EBV-transformed B cells or dermal fibroblasts derived from RCC patients.⁵  Transfection of the rearranged TCR α and β chain genes from the clone into allogeneic T cells was sufficient to transfer the MHC-unrestricted recognition of RCC cells, and a role for TRAIL in target cell recognition was suggested by the observation that recognition of RCC targets was inhibited by an anti-TRAIL antibody.⁵ 

In the current study, Hanada et al screened a cDNA expression library prepared from RCC cells and demonstrated that DR4—also known as TRAIL-receptor 1—was a component of the molecular complex recognized by the RCC-specific T-cell clone, and that DR4 binding its natural ligand TRAIL was involved in target cell recognition. Additional cDNA expression screening revealed that the interaction of the RCC-specific TCR with the TRAIL/DR4 determinant on target cells was strongly enhanced by the interaction of CD2 on the T cells with its cognate ligand CD58 on target cells. Transfection assays revealed that expression of a CD58 cDNA and a truncated cDNA encoding only the membrane-bound extracellular portion of DR4 was sufficient for T-cell recognition, implying that signal transduction by DR4 in target cells was not required. The authors next tested the hypothesis that soluble TRAIL presented by its cognate receptor DR4, in the presence of a stimulatory interaction between CD2 and CD58, was sufficient to trigger T-cell recognition. Indeed, the combination of soluble TRAIL, plate-bound anti-CD2 antibody (to mimic CD58 ligation), and a DR4-Fc fusion protein produced robust T-cell activation. Screening of allogeneic T cells transfected with mutated α and β chain cDNAs encoding receptors with amino acid substitutions in the CDR2 and CDR3 regions of the α and β chains revealed 2 substitutions in the CDR3 loop of the α chain that enhanced both the stimulation generated by soluble TRAIL and plate-bound DR4, as well as T-cell recognition of RCC tumor cells, suggesting that the CDR3α loop of the TCR was directly involved in the recognition of TRAIL/DR4.

In the final act of Hanada et al's enthralling drama, the authors conducted yet another cDNA expression screen to determine why most RCC lines were recognized by the tumor-specific T-cell clone in the absence of exogenous TRAIL. This effort identified matrix metallopeptidase 14 (MMP14), which is expressed on the surface of most RCC lines, as an essential cofactor for T-cell recognition, and prompted the subsequent demonstration that MMP14 could cleave the soluble moiety of TRAIL from membrane-bound TRAIL on T cells. Thus, MMP14 expressed on RCC cells presumably generates soluble TRAIL, which can bind to its ligand, DR4; the TRAIL/DR4 complex can then be recognized by the RCC-specific αβ TCR, with the strength of the interaction enhanced by CD2 binding to CD58 on the RCC target cell (see Figure 7 in the article by Hanada et al, page xxxx).

Among the many questions raised by this elegant study, perhaps the most intriguing deals with the nature of the ligand for the TRAIL/DR4-specific, MHC-independent TCR that mediates its positive selection in the thymus. Is the ligand a conventional peptide/MHC class II complex with which the TCR can cross-react? Or is it some other entity? Solution of the structure of the TCR complexed with TRAIL/DR4 will provide valuable insight into this issue. Equally intriguing is the prospect of using this novel TCR to generate, via gene transfer, MHC-unrestricted, pan-RCC-reactive effector cells for use in adoptive therapy of RCC. The extent to which the TRAIL/DR4 ligand is expressed on normal tissues in vivo will determine the clinical utility of this TCR, and these data are eagerly awaited.