Disease etiology and diagnosis by TCR repertoire analysis goes viral

T-cell receptors are generated somatically during T-cell development through recombination at the tr loci. Rearranged TCR-α and -β chains pair as heterodimers which are subjected to MHC-dependent selection in the thymus in order to seed the peripheral αβ T-cell pool. Thus, the genetic basis of all T-cell-mediated responses is encrypted in the TCR repertoire. The composition of the TCR repertoire is known to incur large changes throughout life, but how TCR diversity varies in health and disease remains poorly understood. However, the tools and methods required to dissect TCR diversity have already begun to reveal invaluable insights into the factors at work to shape the TCR repertoire in the context of disease. Lossius et al. 1 used high-throughput TCR-β chain sequencing to determine the characteristics of the TCR repertoire in multiple sclerosis (MS) patients. Using samples drawn from MS or other inflammatory neurological disorders (OIND) patients, the authors sought to determine the clonal distribution of T cells from cerebrospinal fluid (CSF) and its relationship to T cells from peripheral blood. The CSF TCR-β repertoires from MS and OIND patients overall shared the same features. They displayed similar diversity, compared to blood, with most clonotypes present at a frequency of 0.1% or less. This 0.1% mark was used as the threshold defining expanded sequences throughout the rest of the study. Although the CSF and blood repertoires overlapped partially overall, the most frequent clonotypes were predominantly unique to each compartment. These observations support the idea that TCR-β clonotypes expanding in the CSF are distinct from those expanding in periphery blood. The results of this study mirror those presented in more recent publications 2, 3. Salou et al. 3 examined the TCR-β chain repertoire of three MS patients, using samples from CSF, blood, and central nervous system (CNS) lesion samples. The authors established that the CSF TCR-β repertoire was closer to that of the CNS lesions than the peripheral blood repertoire, again showing that T-cell clones that are predominant in the brain differ from those dominating the blood 3.

Lossius et al. 1 also sought to assess antigen reactivity in CSF-resident clones and found a significant accumulation of TCR-β chain sequences derived from Epstein-Barr virus (EBV)-reactive CD8⁺ T-cell clones. Many of these were private (unique to a patient), although some public (shared between individuals) TCR-β chains were detected, including previously published sequences such as the TCR-β chain of the well-characterized LC13 TCR 4. This is to be expected, particularly for HLA-A2⁺ or HLA-B8⁺ patients in which many public antiviral responses are documented. Although MS is largely suspected to be related to EBV infection 5, the Lossius et al. 1 study was the first attempt to dissect the role of EBV-reactive T cells in this disease based solely on sequencing data. In fact, studies linking antigen specificity to TCR repertoire analysis remain remarkably scarce in the literature. The formation and maintenance of the antigen-specific TCR repertoire are complex processes governed both by genetic and environmental factors, as exemplified by several studies conducted on monozygotic twins 6-10. In the early 1990s, it was first shown that the repertoires of identical twins are a close match at birth but the distribution of TCR clonotypes diverges with age 6. In particular, monozygotic twins can be discordant for autoimmune disease 6, 8, 9, which again, illustrates that the clonal composition of the TCR repertoire cannot be predicted from genetic factors alone and may be a pivotal factor in pathogenesis. Whether the onset and progression of autoimmune disease can be predicted from the repertoire remains unknown, but the integration of high-throughput sequencing technology in the clinical setting means that this hypothesis can now be tested with unprecedented depth.

The relationship between the pathogen-specific TCR repertoire and autoimmunity has long been a matter of interest in the field. With respect to cross-reactivity, molecular mimicry between microbial and self-peptides is often invoked as a potential mechanism leading to the development of autoimmunity. The preproinsulin-specific 1E6 TCR derived from a patient with type 1 diabetes and restricted by the disease-risk HLA-A2 molecule, for instance, can recognize over one million peptides in vitro 11 and a recent structural study indicated that this TCR also recognizes altered peptide ligands mapping to human pathogens 12. 1E6 uses a minimal consensus motif to engage multiple, dissimilar peptides in the context of HLA-A2. This focused binding strategy allows the TCR to tolerate peptide degeneracy outside the binding hotspot 12. ‘Hotspot’ self-reactive TCR binding has been since been documented in the myelin basic protein specific Hy1B11 TCR, which utilizes a minimal binding footprint to contact both self and microbial peptides 13. Thus, while evidence for a role of TCR cross-reactivity in disease development is starting to emerge, monitoring the TCR repertoire in a clinical setting may bring about invaluable insights into the ontogeny of autoimmune disease. Table 1 lists some recent studies attempting to dissect the human TCR repertoire in the context of disease.

Like EBV, human cytomegalovirus (CMV) is known to influence the TCR repertoire 14 and CMV reactivation is linked to immune dysfunction following organ and stem cell transplantation (SCT) 15-17. Suessmuth et al. 18 attributed post-SCT immune dysfunction in CMV-positive recipients to preferential expansion of CMV-specific clones in the effector memory T-cell population, which was found to result in contraction of the overall recipient TCR repertoire 18. CMV has also been previously linked with reduced risk of post-SCT relapse in acute myeloid leukemia patients 19. Thus, dissecting the CMV-specific TCR repertoire in SCT patients may prove beneficial for monitoring transplant outcome, and more generally, for prognostic monitoring of hematological malignancies. For instance, Wu et al. 20 have shown that high-throughput sequencing of TCR-β and TCR-γ chains can detect minimal residual disease in acute lymphoid leukemia better than standard methods involving flow cytometry or quantitative polymerase chain reaction (qPCR) 20.

Public clonotypes derived from virus-specific T-cell clones are particularly appealing as biomarkers because of (i) their high prevalence in the periphery, (ii) the extent to which they are shared in HLA-matched individuals, and (iii) the wealth of information surrounding these clonotypes in the literature. The same would hold true for invariant or semi-invariant TCRs found on the surface of non-classical-MHC-restricted T cells such as invariant natural killer T (iNKT) cells and mucosal-associated invariant T (MAIT) cells. Ancient herpesviruses such as EBV and CMV infect the human population globally and EBV-specific and CMV-specific TCR sequences are commonly found and well characterized 21-26. In other words, while self-reactive TCRs or TCR chains are painfully hard to detect both in tissue and in the periphery, virus-specific TCRs are not. If the latter are readily and reliably detectable in autoimmune disease, as suggested by Lossius et al. and others, they are bound to become important diagnostic tools. Moreover, monitoring the CMV-specific repertoire in solid transplant patients in general, and SCT patients in particular, may well become key to our understanding of the mechanisms leading to transplant acceptance or rejection. Thus, although public and invariant TCRs represent an infinitesimal fraction of the genetic information encrypted in the TCR repertoire, their diagnostic potential in autoimmune disease and hematological pathologies, is certainly worth investigating in future studies.

Conflict of interest

The authors declare no financial or commercial conflict of interest.