Flow-cytometric analysis of rare antigen-specific T cells

Authors


Petra Bacher is employee of Miltenyi Biotec. Alexander Scheffold works as a consultant for Miltenyi Biotec.*Correspondence to: Alexander Scheffold, Charité-Universitätsmedizin Berlin, Charitéplatz 1, 10117 Berlin, Germany. E-mail: alexander.scheffold@charite.de

Abstract

The cytometric enumeration and characterization of antigen-specific lymphocytes, as introduced about 15 years ago, has contributed significantly to our understanding of adaptive immune responses in health and disease. Despite the development of several technologies, allowing to directly or indirectly analyze many aspects of lymphocyte specificity and function, several unresolved issues remain, due to the low frequency of certain antigen-specific lymphocyte subsets and the complexity of T cell antigen recognition. This is especially true for CD4+ conventional as well as regulatory T cells, which bring major contributions to immune protection and pathology. Here we review the current technologies for the analysis of antigen specific T cells within the physiologic T cell repertoire and with a special focus on recent technologies addressing the analysis of rare antigen-specific T cell populations including naive and regulatory T cells. © 2013 International Society for Advancement of Cytometry

Antigen-specific T cells play a central role in mediating specific immune responses as well as in the formation of immunological memory. Information about their frequencies, phenotypes, and functional capacities is essential to estimate the specific immune status of an individual, to understand the mechanisms of protective immunity or immunopathology and to predict immune protection or diagnose immune-related diseases. As a result of the high diversity of the T cell repertoire, allowing to respond to a myriad of different antigens, the frequency of T cells specific for a single peptide-MHC ligand is very low, in particular within the naive repertoire (range 0.2–60 cells/106 naive T cells)[1-19]. But even within the memory repertoire, in the absence of acute infections, specific T cell frequencies in peripheral blood are typically well below 1%. Therefore, the analysis of antigen-specific T cells concerns the detection of rare events, demanding highly specific labeling methods which also allow processing of large cell samples to enable detection of sufficient numbers of the few antigen-specific T cells within the vast majority of nontarget cells.

Due to their low frequencies, antigen-specific T cells have been measured for decades only on the basis of antigen-specific proliferation, for example, via 3H-thymidine incorporation. However, in such assays the phenotypic and functional properties of the reactive cells may be massively altered. Furthermore it is almost impossible to determine the actual frequency and phenotype of reactive cells in the starting cell sample and to exclude bystander proliferation. Therefore in this review, we focus on technologies allowing the direct quantification and characterization of antigen-specific T cells with only minimal in vitro manipulation. Flow cytometry has become the method of choice, since it combines the possibility to simultaneously measure a multitude of parameters for a single cell with the possibility to acquire high event numbers (∼106). Therefore the development of several flow-cytometric methods for the identification of antigen-specific T cells has largely contributed to our current understanding of antigen-specific T cell responses by permitting a detailed phenotypical and functional characterization of each single detected cell. The possibility to measure 20 or more parameters from a single cell using polychromatic cytometry enables to gain maximal information from one measurement, which is essential for small sample sizes or rare subsets. For rare cell analysis the detection limit of flow-cytometric assays can further be decreased by combination with suitable pre-enrichment strategies, such as magnetic cell separation, which allow rapid processing of large cell numbers (>>107). In this way, a sufficient number of target cells can be collected for subsequent flow-cytometric characterization. Such technologies, allow the direct ex vivo detection and characterization of specific T cells for an unbiased knowledge about antigen-specific T cell responses. Here we review the current technologies for direct ex vivo flow-cytometric detection of antigen-specific T cells, including strategies to identify antigen-specific regulatory T cells and to access even extremely rare antigen-specific T cell populations, such as within the naive repertoire. We focus on technologies analyzing T cells within the physiologic repertoire, which are broadly applicable to human T cells. Technologies based on adoptive transfer of traceable numbers of antigen-specific T cells in mice are reviewed elsewhere [20].

Flow Cytometric Methods for the Ex Vivo Detection of Antigen-Specific T Cells

Detection methods can be classified into two categories, which together allow to access T cell specificity and function. Labeling of the specific TCR using recombinant MHC-peptide multimers identifies specific T cells directly according to their specific antigen receptor without restriction to certain functional parameters. This allows unbiased access to the total pool of T cells with a distinct specificity and within a certain affinity range (see below).

The second category employs certain functional parameters as a read-out for T cells, which react to specific antigen challenge. All relevant functions of T cells, such as cytokine release, expression of costimulatory molecules, cytotoxicity and proliferation are accessible via single-cell flow-cytometric assays and for most technologies also, selection markers on the cell surface are available allowing additional isolation of the specific cells.

Direct Labeling of Specific TCRs—MHC-Multimers

MHC-multimers have emerged as an important tool for the identification and characterization of antigen-specific T cells, by direct binding of the T cell receptor to its fluorescently labeled MHC-peptide ligand. Initial problems due to the low binding affinity of the T cell receptor to MHC-peptide monomers [21] have been overcome by the multimerization of peptide-MHC complexes to increase the relative binding avidity [22]. Class I MHC-peptide multimers have widely been used to quantify and characterize antigen-specific CD8+ T cell responses. The construction of class II multimers has initially been more difficult due to differences of the MHC class II structure and TCR affinity, but in the recent years a variety of MHC II multimers for the specific recognition of CD4+ T cells have successfully been generated. In addition CD1 tetramers for detection of lipid reactive T cells including NKT cells have been developed [23-25].

The major limitation of the tetramer technology is that the antigenic epitope has to be characterized in detail, that is, a defined peptide restricted to a particular MHC haplotype. Peptide-exchange technologies allow rapid engineering of numerous different MHC-peptide reagents [26, 27]. Based on this, rapid epitope mapping techniques have been developed, which were successfully used to identify large numbers of new T cell epitopes [28-31]. Furthermore, the use of combinatorial color coded tetramers enabled the simultaneous use of several tetramers to detect a greater number of different antigen-specificities [32, 33]. A distinct combination of fluorophores is used for each T cell specificity, allowing with n fluorophores a theoretical detection of 2n − 1 specificities. So far, this method has been shown to enable the detection of up to 63 different T cell specificities by the use of six different fluorophores [33].

Methods Analyzing T Cell Antigen-Reactivity

The MHC-multimer technology does per se not provide information on the functionality of the detected cells. But in particular, T helper cells can be subdivided in numerous subtypes solely defined by their specific set of effector functions, which are mostly acquired during first antigen encounter and form part of the immunological memory. Thus, information about the functional capacities of antigen-specific T cells and their classification into functionally different subsets is important to determine the quality of an immune response.

The detection of antigen-specific T cells by functional parameters requires the prior in vitro stimulation with the specific antigen. For stimulation, single peptides, proteins or whole antigen lysates can be used, as well as peptide pools, covering the whole sequence of a protein and all possible T cell epitopes. The possibility to use whole proteins and lysates is especially of interest to analyze T cell responses against complex antigens (e.g., whole pathogens or allergens), containing hundreds or thousands of different T cell epitopes. An advantage of activation dependent methods to enumerate antigen-specific T cells is that they are independent of MHC alleles or exact definition of the antigenic epitopes. However, the assays depend on the presence of a sufficient number of functional antigen presenting cells and per definition exclude anergic T cells from the analysis.

Cytokine detection allows identification of functional T cell subsets

The first reports about direct ex vivo enumeration of the frequencies of antigen-specific T cells by flow-cytometry utilized antigen-induced cytokine production [34-36]. Cytokine expression is mostly confined to TCR activated T cells and it is transient, typically measured within 4–12 h, although within this time window different cytokines might have different kinetics. Cytokines can be detected either intracellularly when cells are simultaneously incubated with secretion inhibitors like Brefeldin A or Monensin [37, 38] or on the cell surface by retention of the secreted cytokine on the surface of the secreting cells via a capture matrix [34, 39]. The latter method has the advantage that surface retained cytokines enable the detection of live cytokine secreting cells as well as the enrichment via magnetic cell sorting. However, almost all cytokines are restricted to certain T cell subsets and in particular naive T cells produce only few cytokines upon stimulation. Therefore the enumeration of antigen-specific T cell frequencies solely based on cytokine production may be incomplete and has to be handled with care. On the other hand, the enrichment of highly specialized subsets can have significant advantages, for example, purified virus-specific IFN-γ secreting T cells are ideally suited for adoptive T cell therapy of viral infections [40-43].

Activation markers identify the total pool of reactive T cells

An alternative approach for the direct visualization of antigen-specific T cells by flow cytometry is the detection of activation markers, which are upregulated on the T cell surface upon antigen-specific triggering. Expression of some of these markers is independent of other effector functions like cytokine secretion or cytotoxicity, which may be restricted to the differentiation state of the T cells (e.g., naive, central memory, effector memory) or certain T cell subsets (e.g., Th1, Th2, Th17, etc.). Therefore activation markers permit the comprehensive characterization of the total pool of specific T cells against a given antigen, irrespective of functional specialization, MHC allele or exact definition of the antigenic epitope, as outlined above. A prerequisite for high specificity is that the marker expression should solely depend on T cell receptor triggering and expression should also be transient. A number of potential activation markers have been proposed, including CD69, CD25, CD71, HLA-DR, CD134 (OX40), CRTAM, CD137 (4-1BB), CD154 (CD40L) [44-55]. However, a limitation for many of these markers is their sensitivity to bystander activation (CD69, CD25), their constitutive or exclusive expression on specialized T cell subset (CD69, CD25, CRTAM), or their late upregulation following in vitro stimulation (HLA-DR, CD134, CD71), which affects their usability for accurate enumeration of antigen-specific T cells occurring at low frequencies (<<1%).

One widely used activation marker for CD4+ and CD8+ T cells is CD69, which is one of the earliest markers (3–15 h; [44]) expressed on activated T cells, as well as B cells or NK cells [56-59]. However, as CD69 background expression is also found in variable amount on nonstimulated T cells [49], and CD69 upregulation is not solely dependent on T cell receptor triggering [60, 61], analysis of CD69 expression alone may strongly overestimate the frequencies of antigen-specific T cells. In contrast, in combination with other early activation markers (e.g., CD154 and CD137, see below) or cytokine expression, the analysis of CD69 coexpression as a second parameter is a useful tool to increase the sensitivity and optical discrimination of rare cells in flow-cytometry data.

For CD4+ T cells, CD154 (CD40L), a member of the TNF superfamily, has been shown to be a reliable functional marker for the detection of antigen-activated T cells [45-47]. As the central mediator of T cell help CD154 is expressed by virtually all functional activated CD4+ T cells irrespective of their differentiation status, as well as by a subset of CD8+ T cells. Another technically important feature of CD154 is its extremely low ex vivo background expression, which allows a highly specific detection of antigen-induced CD154 expression [62]. This is most probably due to its rapid internalization and degradation following interaction with its receptor CD40 expressed on antigen-presenting cells. CD154 expression can readily be assessed either intracellularly by adding secretion inhibitors like Brefeldin A or Monensin to the culture or on the surface by blocking the interaction with its ligand through addition of an anti-CD40 mAB. As CD154 upregulation occurs fast (4–12 h) following antigen encounter, its detection can easily be combined with staining for cytokines and phenotypic markers.

Another well described marker with sufficiently high specificity is CD137 (4-1BB), a member of the TNFR superfamily, which has been shown to be expressed on antigen-activated CD4+ and CD8+ T cells as well as γδ+ T cells, following 16–24 h of stimulation [51-53]. Recently, CD137 has also been shown to be expressed on antigen-activated CD4+Foxp+ regulatory T cells (Treg, see below) [63]. Since both, CD154 and CD137 are expressed on the surface of the antigen-specific T cells, enrichment and analysis of living cells is possible allowing further functional characterization or clinical-scale enrichment of antigen-specific T cells for adoptive therapies.

Accessing Rare Antigen-Specific T Cells

Despite its high sensitivity, standard flow-cytometry is limited by the number of events, which can be acquired (typically ∼106 cells) as well as the staining background, which is typically between 0.01 and 0.1%. This usually restricts the analysis of un-manipulated samples to populations occurring at frequencies >0.01–0.1%. Therefore, the technologies described above have mainly be used for the detection of pathogen specific memory T cells, although the frequency of pathogen-specific T cells can vary widely, depending on the nature of the pathogen, the status of the immune response and the persistence or clearance of the pathogen. In absence of an acute infection, frequencies of antigen-specific memory cells are typically in the range of one cell within 100 to 105. However, many functionally important T cell populations occur at much lower frequencies and cannot be accessed without additional measures. This applies in general for the naive compartment, as well as the CD4+ T cell repertoire and in particular for the small population of regulatory T cells. Furthermore T cells specific for nonpathogen derived antigens, that is, auto-antigens, tumor antigens, environmental antigens (allergens, toxins, food) or neo-antigens are typically below 0.01% (Fig. 1). Methods to enumerate such rare populations must therefore allow to process large cell numbers, to collect sufficient target events for statistical relevant analyses and at the same time prevent accumulation of background events.

Figure 1.

Frequencies of antigen-specific CD4+ and CD8+ T cells within human peripheral T cell repertoire determined directly by ex vivo enrichment methods. Dashed lines indicated the detection limit range of standard flow cytometric counting and magnetic enrichment methods, respectively.

In Vitro Expansion of Antigen-Specific T Cells to Facilitate Detection

One solution to the rare cell problem is to increase the number of antigen-specific T cells by in vitro expansion methods. Numerous approaches have been published using cultivation of un-separated PBMC with the particular antigen for 1–2 weeks allowing the specific T cells to grow preferentially. However, the expansion of a single T cell is affected by a number of hardly predictable parameters such as its functional status (e.g., naive, memory, anergic), the presence of other reactive cells (bystander activation), or the relation between the number of cell divisions and cell death. Therefore, it is difficult to extrapolate from the frequencies obtained after prolonged in vitro culture to the frequencies of cells in the original samples, even if the proliferative history of the cells has been visualized, for example, by use of cell proliferation dyes. Similarly, the phenotype and function of the expanded cells may significantly be altered by the culture conditions.

The group of Sallusto recently described the use of libraries of polyclonal expanded CD4+ T cells, to determine the frequencies of antigen-specific T cells within the human naive and memory repertoire [64]. Each library consists of several hundred single cultures, each seeded with 2,000 T cells cells/well, so that at the most one specific T cell/well could be expected. The cells are polyclonally expanded while maintaining the TCR diversity and restimulated with one or several antigens in the presence of autologous monocytes. The original frequency of antigen-specific T cells can then be calculated based on the number of proliferating cultures and the number of input cells. Using this approach, the authors could detect frequencies of antigen-specific naive CD4+ T cells for protective antigen from Bacillus anthracis between 10 and 26 per 106 naive CD4+ T cells and for KLH between 10 and 70 cells per 106 naive CD4+ T cells. An advantage of this method is that it enables the enumeration of antigen-specific T cells specific for naturally processed antigens. By presorting of different T cell subpopulations for set-up of the libraries, antigen specific T cell frequencies from various T cell compartments can be determined and compared. Once generated, the libraries can be used for screening of several antigens and for further functional analyses such as TCR affinity determination. However, the technique requires maintaining several hundreds of individual cultures for several weeks. Furthermore, a bias through selective outgrowth or loss of certain T cell clones cannot be excluded, especially for mixed starting populations, for example, different memory subsets, due to differential expansion and survival potential. Finally, the detection limit is determined by the number of individual cell cultures per library and the number of input cells per culture. For a frequency of one cell within 105 only 10 positive wells are detected within 500 cultures (i.e., 2,000 cells/culture, 106 cells in total). This basically restricts the analysis to a frequency window from 1 cell within 2,000 (all cultures positive) to maximally one cell within 105 (10 cultures positive). For example the authors fail to detect the rare specific T cells against neo-antigens within the memory T cell compartment, which have been previously described [65] and can be readily detected by other approaches [62, 66].

Antigen-Specific Pre-Enrichment

Magnetic enrichment to collect few target cells from large cell samples

An alternative approach to increase the number of rare antigen-specific T cell frequencies for proper cytometric analysis utilizes quantitative pre-enrichment of target cells via magnetic cell separation, which allows rapid processing of large cell samples (106–109) [67-70]. In this way, all labeled antigen-specific T cells (e.g., peptide-MHC tetramer, activation marker, surface captured cytokines) from a large sample are retained on the column and can subsequently be conveniently analyzed by flow-cytometry or used for further functional studies or expansion. A prerequisite for the magnetic enrichment of antigen-specific T cells is a high specificity of the sorting marker to reliable identify the few antigen-specific T cells within the vast number of irrelevant cells. As the eluted fraction still contains significant frequencies of nontarget cells, additional exclusion criteria, such as staining of non-T cell lineage cells, doublets and dead cells are required to increase sensitivity and specificity. In combination with multi-parameter flow cytometry, this technology allows the direct ex vivo detection and in depth characterization of few antigen specific T cells from large sample sizes. The magnetic enrichment of antigen-specific cells leads to a dramatic increase of sensitivity, which is mainly limited by the number of available input cells, and gives important new insights into previously undetectable parts of the T cell repertoire, including even the naive compartment. The general characteristics of different enrichment methods for the detection of antigen-specific T cells are summarized in Table 1.

Table 1. Enrichment methods for the detection of rare antigen-specific T cells
MethodAdvantagesDisadvantagesDetectable T cell populations
Direct detection   
MHC-multimer
  • Activation independent
  • High specificity
  • Knowledge/availability of MHC and epitope required
  • Restricted to single epitope specificities
  • Does not reveal functional status of the cells
  • Detection of low-affinity cells difficult
Naive, memory, Treg
Activation dependent
  • No MHC/epitope restriction
  • Reveal functional status
  • Dependent on T cell activation
 
Cytokine secretion assay
  • Selective isolation of distinct cytokine producing subsets
  • Restricted to few selected cytokine producers
  • Laborious
  • Risk of contaminating cells through cross-feeding
Memory
CD154 (6 h)
  • Detection of the whole antigen-specific CD4+ T cell response
  • Fast kinetics
  • Compatible with cytokine analysis
  • Mainly restricted to CD4+ T cells
Naive, memory
CD137 (16 h)
  • Detection of the whole antigen-specific CD4+, CD8+ T cell response
  • Not compatible with cytokine analysis
  • No differentiation between Treg and conventional CD4+ T cells (due to 16 h stimulation)
Naive, memory
CD154/CD137 (6 h)
  • Detection of the whole antigen-specific CD4+ conventional and regulatory T cell response
  • No CD8+ T cells
Naive, memory, Treg

Enrichment of cytokine secreting cells

The enrichment of rare cytokine producing T cells, enabled by the cytokine secretion assay was the first report utilizing magnetic enrichment for the analysis of rare antigen-specific T cells undetectable via conventional flow-cytometry [34].

This has been extended to several other rare cytokine producing T cell subpopulations. For example, tetanus-specific IFN-γ+ and IL-4+ CD4+ T cells or C. albicans-specific IL-17+ CD4+ T cells were isolated with frequencies of approximately one cell within 104–105 PBMC [34, 71] and auto-reactive IFN-γ+ and IL-4+ CD4+ T cells in pemphigus vulgaris patients could be detected with frequencies of 3–40 cells per 105 PBMC and less than 10 cells per 105 PBMC in healthy individuals [72]. In addition, rare allergen-specific IL-4, IFN-γ, and IL-10 producers have been isolated from allergic patients and healthy individuals with frequencies of 1–10 cells within 104 CD4+ T cells [73]. The direct visualization of allergen-specific cytokine secreting cells revealed a higher frequency of IL-10 producing cells in healthy individuals, whereas in allergic patients IL-4 producers were dominant indicating that the balance between IL-10 and IL-4 producing cells might be decisive in the development of allergy [73].

Enrichment of rare IFN-γ producing cells has also been instrumental for clinical adoptive immunotherapy, for example, clinical-scale isolation of rare virus-specific T cells [40-43].

However, as outlined above cytokine secretion is heterogeneous and restricted to certain T cell subsets, which often represent only a small fraction of the total antigen-specific T cells. Although the cytokine secretion assay allows simultaneous detection of more than one cytokine, it is limited to maximally two or three, since the labeling of an individual cell with different catch matrices for the respective cytokines reduces the labeling intensity, which precludes their proper identification as well as magnetic enrichment.

Tetramer enrichment

MHC tetramers were also used to detect and isolate rare antigen-experienced CD4+ and CD8+ memory T cells of humans and mice [26, 74-82]. The sensitivity of the tetramer-based enrichment technology was further extended to the identification and first direct enumeration of the extremely rare antigen-specific CD4+ and CD8+ T cells within the naive T cell repertoire. In the murine system, naive CD8+ T cells specific for epitopes of ovalbumin or various viruses could be detected with varying frequencies from 15 to 1,500 cells per mouse [7-14, 18]. The peptide-MHC class II tetramer based enrichment revealed frequencies of naive CD4+ T cells specific for epitopes within ovalbumin, Salmonella typhimurium, Listeria monocytogenes, I-E alpha chain or the interphotoreceptor retinoid binding protein (IRBP) from 20–260 cells per mouse [1-6]. Given an average number of 2.5 × 107 naive CD8+ and 3.5 × 107 naive CD4+ T cells per mouse [83], the identified frequencies of T cells for an individual peptide-MHC ligand range from 0.6 to 60 cells per 106 naive CD8+ T cells and 0.6–7.4 cells per 106 naive CD4+ T cells. Also in humans, antigen-specific naive T cells were enumerated by tetramer-based enrichment with surprisingly similar frequencies: CD4+ T cells specific for a peptide of influenza hemagglutinin and three different epitopes of Bacillus anthracis ranged between 0.2 and 10 cells per 106 naive CD4+ T cells [2, 16], which is in the same range as estimated by the library method [64]. Within the CD8+ T cell repertoire antigen-specific naive T cells for epitopes of different viruses and the auto-antigen NY-ESO1 were analyzed, with frequencies of 0.5–30 per 106 naive CD8+ T cells [15, 17, 19].

An interesting finding of these studies was that the frequencies of naive precursor T cells for different antigens can vary widely, but showed limited inter-individual variation. This is especially surprising in humans, as HLA haplotypes are highly diverse. Furthermore, the size of the naive T cell population specific for a given antigen could be linked to immune dominance [13, 84], as it was shown that naive precursor frequencies correlate with the size of the memory T cell response [4, 10, 13, 16, 19, 84].

Only recently a completely different mechanism for shaping the human memory repertoire has been highlighted by Davis and coworkers [66]. They used highly sensitive HLA class II tetramer enrichment to show that the peripheral blood of human individuals which have not been infected with a certain virus before contain significant frequencies of bona fide memory T cells specific for the very same virus, suggesting that cross-reactivity during life contributes significantly to shaping of the immune repertoire. Also our own data, using ex vivo enrichment of CD154 expressing cells following auto-antigen stimulation suggest, that a significant proportion of the self-reactive CD4+ T cell pool in healthy individuals has a memory phenotype, which may result from cross-reactivity to other antigens ([62] and see below).

Tetramer based enrichment has also been used to analyze rare allergen-specific T cells as potential correlates for disease status or efficacy of immunotherapy. Because of their low frequencies, previous analyses of allergen specific T cells using MHC-multimers were mainly performed after in vitro expansion, which can alter the phenotype and functional capacities of the cells, as already discussed above. Direct ex vivo tetramer enrichment of allergen-specific CD4+ T cells has recently been used to analyze specific T cells against epitopes from cat, peanut, alder, and birch allergens [85-87]. Allergen-specific T cells were barely detected in peripheral blood of nonatopic individuals with frequencies of about one to five cells within 106 CD4+ T cells. In contrast, average–frequencies in allergic patients ranged from 7–500 cells within 106 CD4+ T cells. The differences in the frequencies of allergen-tetramer positive cells might reflect, that allergen specific T cells are present at lower frequencies in nonallergic subjects. Alternatively, a lower avidity of allergen-reactive T cells in healthy individuals has been suggested [88], and tetramers may be limited in their ability to detect low-affinity T cells [89]. Although target proteins and T cell epitopes are defined for many allergic diseases, the tetramer technology restricts the analysis to a few pre-selected epitopes and MHC haplotypes and might therefore underestimate the frequencies of allergen specific T cells in healthy donors. In fact, our own preliminary results, analyzing the total allergen-reactive T cell pool using CD154 enrichment after stimulation with whole proteins or allergen extracts (see below), indicate no difference in specific T cell numbers between healthy and allergic subjects but a modulation of memory T cell subsets, as well as a strong shift towards a Th2 phenotype (Bacher et al., unpublished).

Rare antigen-specific CD4+ T cell analysis using CD154 enrichment

The advantages of CD154 as a marker for the total pool of antigen-reactive T cells were discussed above. We recently demonstrated that CD154 enrichment can also be utilized to strongly increase the sensitivity of detection for rare cells (10−5–10−6 largely depending on the number of input cells) [62]. Using this approach, a broad functional heterogeneity of virus-, bacteria- or fungus-specific CD4+ T cells was visualized, demonstrating that many functional relevant cytokines occur only in very low frequencies within the total antigen specific T cell pool, undetectable without pre-enrichment methods (1–10% of CD154+ cells, which corresponds to 0.0001–0.1% within the CD4+ T cell repertoire). Despite this heterogeneity, these “cytokine micro-patterns” were characteristic for a specific pathogen, indicating a pathogen-specific immune modulation, which cannot be classified according to the simplified Th1, Th2, Th17 lineage differentiation scheme. The increased sensitivity through the magnetic enrichment was further used to directly analyze the full repertoire of KLH-, HIV- or several auto-antigen-specific CD4+ T cells in the peripheral blood of healthy donors. The frequencies of these cells were in the range of 1 cell within 104 to 106, which is similar to the results obtained using the “T cell library approach” [64]. However, we could furthermore demonstrate that about on third of the identified neo- or auto-antigen-specific CD154+ T cells in healthy donors had a memory phenotype, suggesting that these cells are the result of priming by cross-reactivity, which extends the findings by Mark Davis group for virus-specific T cells using tetramer enrichment [66].

Antigen-Specific Regulatory T Cells

A further challenge is the detection and enumeration of antigen-specific CD4+Foxp3+ regulatory T cells (Treg). Despite their importance for the regulation of immune responses and the maintenance of tolerance, exact information of antigen-specificity and the frequency of antigen-specific Treg is lacking. Fragmentary knowledge of specific target antigens or even antigenic peptides limits the applicability of MHC class II multimers. In addition, less is known about the affinity of Treg cells to their peptide-MHC ligands, which might differ from conventional T cells and may be below the detection threshold of MHC class II multimers. Treg also lack most of the effector functions typically used for the analysis of conventional T cells, such as cytokine production or in vitro proliferation. Finally, sensitive analysis tools are particularly relevant for antigen-specific Treg, since they usually represent a subset of only 5–10% within the circulating human CD4+ T cell compartment but have similar TCR diversity, which means that antigen-specific Treg are extremely rare. Therefore, enrichment strategies are essential for their reliable detection in nonmanipulated biological samples.

Tetramer Detection of Treg Cells

MHC class II tetramers have been successfully used to analyze antigen-specific Treg cells in studies of autoimmunity, allergy, transplantation or pathogen-specific immune responses. A recent study in mice addressed the question of thymic selection impact on Foxp3+ and Foxp3 self-peptide specific CD4+ T cells, using tetramer enrichment [90]. Self-peptide specific CD4+ T cells were not deleted in mice that ubiquitously expressed the antigen, but both Foxp3 and Foxp3+ subsets were reduced compared to wildtype mice. However, the remaining T cells were enriched for Foxp3+ cells, indicating a higher resistance of Foxp3+ T cells to negative selection. Another important finding of this study was that Treg cells specific for foreign antigens could be detected in the CD4+ T cell repertoire of mice, that were never exposed to the antigen before [90, 91]. The frequencies of Foxp3+ cells specific for epitopes of I-E α-chain, Lymphocytic choriomeningitis virus, Listeria monocytogenes or ovalbumin ranged from 7.5 to 12% of the respective specificities in the naive T cell repertoire [90].

In humans MHC class II tetramers were used to identify antigen-specific Treg cells for birch, alder, and timothy grass allergens, as well as epitopes from Streptococcus pneumonia, Varicella zoster, Mycobacterium tuberculosis, Hepatitis C Virus, Influenza, HIV or the islet specific antigens GAD and IGRP [87, 92-99]. However, due to their low frequencies, in most of these studies antigen-specific Treg cells were previously expanded for several days, by stimulation of bulk cultures with the respective antigenic epitopes. Under these conditions, differentiation between Treg and conventional T cells is difficult, since Treg specific markers such as CD25 and Foxp3 or lack of CD127 expression are also acquired by non Treg cells upon activation [100]. This further underlines the importance to analyze and enumerate antigen-specific Treg cells directly ex vivo.

Recent studies using MHC II tetramers have demonstrated directly ex vivo the existence of human regulatory T cells specific for foreign antigens. Foxp3+ Treg cells specific for a peptide of the major birch pollen allergen Bet v1 were detected in human tonsils with frequencies of approximately one cell within 1,000 CD4+ T cells, which corresponded to about 30% of the total tetramer positive population [93]. However, in peripheral blood no specific Treg cells could be detected. Similarly, tetramer positive Streptococcus pneumonia specific Foxp3+ Tregs could be detected in human tonsils, with frequencies ranging from 0.53 to 1.96% among the total tonsillar Treg cell population, whereas among peripheral blood Treg cells no tetramer positive cells were found [96]. The different Treg frequencies in tonsils versus peripheral blood argue for a selective accumulation of foreign antigen-specific Treg cells at the site of antigen encounter. Also after cutaneous Varicella zoster or Mycobacterium tuberculosis challenge, human antigen-specific Tregs were identified by specific tetramers at the site of antigen-challenge in the skin [97]. However, when tetramer enrichment was used, antigen-specific Tregs against Hepatitis C virus, Influenza and alder allergen could directly be detected in human peripheral blood with frequencies of ∼5–15% of the total tetramer positive population [87, 95], which is surprisingly similar to the 7.5–12% Treg cells within the foreign peptide-specific naive T cell repertoire detected in mouse [90]. These data in fact suggest that Treg are part of the natural repertoire against foreign antigens and that specific Treg coexpand together with conventional T cells during an immune response most likely to control or prevent excessive immune reactions.

Treg Specific Activation Markers

Although the typical markers of TCR-activated conventional T cells are lacking on Treg cells, several Treg specific activation-markers have been introduced that allow the isolation of activated Treg cells after in vitro stimulation, including CD121a/b, latency-associated peptide (LAP) and GARP (LRRC32) [101-103]. But so far none of these markers has been evaluated for an antigen-specific analysis of Treg cells. However, only recently the activation markers CD154 and CD137 have been reported to be upregulated on Treg cells following short term in vitro antigen stimulation [63, 104]. Schönbrunn et al. show that the kinetics of activation marker expression differ between Treg and conventional T cells, that is after 6 h stimulation only Tregs express CD137 whereas conventional T cells require >12 h of stimulation [63]. In this study, also the relation between CD154 and CD137 expression was analyzed. CD137 is expressed by all activated Tregs, whereas CD154 expression is restricted to a subset of Treg cells which actually seem to represent a population with instable Foxp3 expression as shown by methylation analysis of the Foxp3 promotor region. Thus the combined analysis of CD137 and CD154 following short-term (6 h) stimulation might be optimal to detect in parallel conventional T cells and Tregs reacting against the same antigen and allows even to discriminate between subsets with stable (CD137+CD154) or unstable (CD137+ CD154+) Foxp3 expression. Schönbrunn et al. used the technology to directly sort highly suppressive allo-reactive Tregs out of PBMC stimulated with allogenic APCs.

Since both markers can be accessed on the cell surface they can also be used in combination with a magnetic pre-enrichment of the activation marker positive cells [62]. This approach thus enables the detection of rare antigen-specific cells with high sensitivity, but without limitations of TCR affinities or restriction to certain epitopes and could therefore be extended to any antigen of interest. Our own preliminary results using a combined CD137/CD154 enrichment indeed show that antigen-specific Treg can be detected with high sensitivity (Fig. 2) and we are currently using this approach to characterize antigen-specific Treg cells from human peripheral blood directed against environmental antigens, as well as auto-antigens.

Figure 2.

Detection of antigen-specific conventional CD4+ T cells and regulatory T cells following combined magnetic enrichment of CD154+ and CD137+ cells. PBMC of a healthy human volunteer were stimulated for 6 h with A. fumigatus lysate. CD154+ and CD137+ CD4+ T cells were isolated by magnetic enrichment and counterstained for CD25 and CD127 expression. For an optimal detection of CD4+ CD154+ and CD137+ events, cell aggregates (scatter area versus scatter height), dead cells and nontarget cells (CD8+ CD14+, CD20+, dump) were excluded and cells were gated on CD4+ lymphocytes. Numbers in brackets indicated cell count within 2.5 × 105 analyzed PBMC before enrichment and following enrichment from 1 × 107 PBMC. FSC, forward scatter; SSC side scatter; LD, live/dead; PE, phycoerythrin; APC, allophycocyanin. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Conclusion

Recent developments which significantly increase the sensitivity for cytometric detection of antigen-specific T cells and also enable the detection of antigen-specific Tregs, now allowing to analyze the full CD8+ as well as CD4+ T cell repertoire against most pathogens, environmental antigens, and tumor- or auto-antigens, even in the unexposed host. In particular, the possibility to simultaneously analyze T cell responses in the naive, effector as well as regulatory T cell compartment will help to define their role within the immune network and to deduce their lineage relationship. Understanding of the full antigen-specific repertoire will significantly improve diagnostic and prognostic as well as therapeutic intervention strategies.

Although the various technologies yield comparable results in terms of cell frequencies they highlight different aspects of T cell specificity and function and therefore integrative approaches are mandatory. The main limitation seems to be the availability of sufficient sample material, especially from patients, which allows to collect a suitable number of target cells. The small number of target cells also makes high demands on downstream analysis tools and therefore multi-parameter analysis tools such as polychromatic flow-cytometry will be essential to extract maximal information from the few available cells.

Ancillary