Flexibility for specificity

See allHide authors and affiliations

Science  23 Jan 2015:
Vol. 347, Issue 6220, pp. 371-372
DOI: 10.1126/science.aaa5082

Different types of T lymphocytes play a key role in many immune responses, such as killing virally infected or cancerous cells directly, inducing high-affinity antibody responses in B cells, and increasing or decreasing responses from other immune cells. This multiplicity of roles may relate to their recognition properties, which are very difficult to evade. Moreover, cells are very diverse—for CD4+ T cells alone, there are at least six distinct subtypes. This raises the question of just how these different T cells are produced. Early evidence indicated that the type of T cell that dominates the response was dependent on the type of pathogen and route of entry (1, 2). However, over the past several years, more and more flexibility has been observed in a T cell's phenotype (3, 4). On page 400 of this issue, Becattini et al. (5) show that this flexibility is more the rule rather than the exception.

Most T cells use antibody-like T cell receptors (TCRs) on their cell surface to recognize degraded bits of proteins (peptides) or lipids that are bound to molecules of the major histocompatibility complex (MHC) expressed on the surface of various cell types. These MHC molecules are much like the people in Hollywood who go through the garbage cans of celebrities, looking for interesting bits of information to display to the tabloid press. But MHC molecules cannot tell whether the fragments they bind are from endogenous proteins of the cell; that is the job of T cells, with their diverse TCRs. The functional beauty of this system, and perhaps the reason for its centrality in so much of adaptive immunity, is that everything pertaining to a cell or a pathogen is eventually degraded, making it much harder for pathogens to evade detection. By contrast, antibodies must recognize intact antigens and thus are often diverted from the most important bits. In particular, MHC molecules bind small peptides—typically 8 to 10 amino acids—just enough information to be fairly unique, making it easier for T cells to discern what belongs in the body (“self”) and what doesn't. Although this rationale is somewhat speculative, the diversity of T cell types is very clear.

Becattini et al. used two pathogens (Candida albicans and Mycobacterium tuberculosis) and a protein (tetanus toxin with the adjuvant alum) to show that a variety of different T cell types can be derived from a single initial T cell clone. Because TCRs do not mutate, unlike antibody genes, it is relatively easy to follow a particular T cell clone through its various divisions, and (in this case) the different functional properties of the daughter clones, just by sequencing the TCRs expressed by the clonal population and correlating sequences with phenotypic information.

One cell, multiple fates.

A naïve human T cell primed with a pathogen in vitro gives rise to multiple T cell subtypes through a proposed two-phase process. In the first phase, specific T cells are stimulated to proliferate regardless of their phenotype. In the second phase, there is a selective expansion of T cells with the most appropriate functional properties, and multiple types can arise from the same original T cell clone. This model may explain how T cells acquire their specialized responses to infection and vaccines.


Becattini et al. focused on four CD4+ T cell types: T helper 1 (TH1) cells, which produce the cytokine interferon IFN-γ to combat intracellular bacteria and viruses and activate macrophages; TH2 cells, which make interleukins IL-4, IL-5, and IL-13 and help control parasites and activate eosinophils; a recently discovered TH1* type (also known as TH1/17) that makes some IL-17 in addition to IFN-γ; and TH17 cells, which produce IL-17 and IL-22, mobilize neutrophils, and combat fungi and extracellular bacteria. The authors took peripheral blood cells from up to five human donors, loaded them with carboxyfluorescein succinimidyl ester (CFSE, a fluorescent dye that allows one to track the number of cell divisions), and stimulated the cells with the fungus C. albicans. Memory T cells were recovered in bulk or isolated from each of the four different T cell populations surveyed, with the criterion of low CFSE staining to collect only those T cells that had divided in response to the pathogen. The TCRβ chains were then sequenced in bulk for each population, using high-throughput techniques. As expected for this pathogen, TH17 cells were the most common, but TH1* cells also were abundant. Surprisingly, all four cell types were represented and were relatively equally diverse (595 to 976 unique TCRβ sequences in each population), indicating that at least initially, T cells of all predispositions were drawn upon. Most important, there was widespread sharing of TCRβ sequences in the different populations; in some cases, the same sequence occurred in all four T cell types. Subsequently, T cell cloning (deriving a population of T cells from a single member of each T cell population) confirmed these results (each cloned population had a TH1, TH2, TH1*, or TH17 phenotype), as did determining that the same TCR sequences were present in the TCRs across these different cloned T cell populations. The same bulk approach was also used to analyze responses to M. tuberculosis and tetanus toxin, the former an intracellular mycobacterium and the latter a single protein. In the case of M. tuberculosis, the data were similar with respect to the initial diversity of the different TCR repertoires, but there was much less sharing of sequences between T cell types. For tetanus toxin, there was widespread sharing of TCR sequences, even more than observed with C. albicans.

The extensive analysis of three different antigens makes a strong case that the recruitment of at least these types of CD4+ T cells is relatively random at first, but that specific functional cells can be expanded preferentially by some process of selection and interaction with the pathogen or antigenic challenge (see the figure). That this principle is likely to extend to other CD4+ T cell types and CD8+ T cells is suggested by an earlier study that showed CD8+ T cells specific for a particular viral epitope could express more than 100 different combinations of cytokines or chemokines (6). Although the study did not involve TCR sequencing, it does show that T cell responses can be much more diverse functionally than previously thought.

A question yet to be explored concerns how particular types of T cells are favored by pathogens (or by adjuvants, as is likely the case with the tetanus toxin exposure). Knowing the answer could help in the design of vaccines that boost the types of T cells that can do the most good. Another question is how flexible these T cell responses are: How likely are they to change with time, and under what influences? Data show that there is a delicate balance that can tip back and forth between TH1- and TH2-type cells (7). Other results show changes in the cytokine secretion of single CD4+ T cells with time (8). And there is improved technology for deriving both TCR sequences and phenotypic information from single T cells (9), which will make studies like this much easier, and perhaps unravel even further the degree of heterogeneity in T cell responses to infection and vaccines.

References and Notes

  1. Acknowledgments: Supported by the Howard Hughes Medical Institute, NIH, and the Bill and Melinda Gates Foundation.
View Abstract

Stay Connected to Science

Navigate This Article