The immune system provides protection from various infections and tumors, but can also mediate allergies, autoimmunity and transplant rejection. Over the last decades, it has become evident that in each of these cases different types of effector T cell classes play a role, the primary ones being Th1, Th2, and Th17 [1
]. Th1 cells mediate their effector functions via the secretion of interferon gamma (IFN-γ) that inhibits viral replication in infected cells, and upregulates major histocompability complex (MHC) antigen expression on such cells priming them for T cell recognition. Th2 cells secrete IL-4 that is involved in regulating antibody-mediated immunity by controlling immune globulin class-switching. A sublineage of Th2 cells is involved in anti-parasite defense via the secretion of IL-5. Finally, the Th17 cell type mediates delayed type hypersensitivity (DTH) by secreting IL-17 [2
]. In DTH, macrophages are recruited and activated to constitute the primary line of defense against intracellular pathogens.
The Th1, Th2 and Th17 lineages emerge through instructed differentiation. Naïve T cells are uncommitted and can develop into either of these cell types [3
]. During the primary immune response, when naïve T cells first encounter antigen, they start to proliferate, and dependent on the cytokine environment in which this reaction occurs they also engage in differentiation. In the presence of IL-12, Th1 cells emerge while in the presence of IL-4 Th2 cell development occurs [4
]. If IL-6 and TGFβ prevail in the microenvironment, Th17 cells emerge [7
]. Once Th1/Th2/Th17 differentiation is completed, which happens within the first 10 days of the primary immune response [8
], the cytokine expression profiles of these cells are firmly imprinted and mutually exclusive [9
]. Upon antigen re-encounter, Th1 cells will secrete IFN-γ (but no IL-4 or IL-17), Th2 cells will produce IL-4 or IL-5, but no IFN-γ or IL-17, and Th17 cells IL-17 in the absence of these other cytokines [10
]. In this way, dependent on the ratios of antigen-specific Th1/Th2 and Th17 generated, the T cell system can elicit highly specific effector mechanisms on the site of antigen re-encounter.
Beyond mounting a T cell response per se, the magnitude of clonal expansion within the Th1/Th2/Th17 lineages is a primary factor defining whether protective immunity develops. Engaging, for example, Th2 immunity when Th1 or Th17 would be the adequate class of response can be deleterious for the host [12
]. Also, allergies and autoimmunity can be viewed primarily as the consequence of engaging an inappropriate T cell effector class [13
]. Therefore, our abilities to unambiguously define frequencies of Th1, Th2 and Th17 cells is at the heart of any immune monitoring effort aiming at better understanding immune-mediated processes, whether beneficial to the host, or harmful. The inclusion of cytokines other than IFN-γ into immune monitoring efforts can also be critical for identifying disease activity states. For example, it has been shown that IFN-γ ELISPOT (enzyme-linked immunospot) assays alone are not suited to distinguish between the active and latent form of tuberculosis infection in children, however, additional IL-2 testing helps to make this discrimination [14
Detecting antigen-specific Th1/Th2 and Th17 cells relies on measuring the lead cytokines these cells produce following antigen exposure. Resting Th1, Th2 or Th17 cells, like memory cells in general, do not constitutively secrete cytokine but are preoccupied with recirculation in the body while seeking antigen [15
]. Such cells are the typical lymphocytes seen in peripheral blood with a prevalent nucleus and underdeveloped cytoplasm and endoplasmic reticulum (ER) enabling protein synthesis. Upon re-encounter with antigen, these memory T cells undergo blast transformation during which their ER is re-developed, and they gain the ability to express quantities of cytokine. For Th1 cells, it is well established that it takes 12–24 h after antigen re-encounter before IFN-γ production reaches its peak [16
]. In mice induction of IL-4, IL-5 and IL-17 by the corresponding T cell subset also peaks within 24 h [17
]. In humans, however, there have been occasional observations of delayed IL-4, IL-5 and IL-17 secretion kinetics by memory T cells [18
], but the detection of these effector T cell classes in PBMC has mostly been unsuccessful, likely because of the low frequency of such T cells, and possibly because too short antigen-stimulation cultures have been used. A systematic study to this extent is missing so far, and is presented in this report. The differences between mouse and humans might be species dependent, but might also reflect on the time elapsed since the last antigen encounter of the memory cells [19
]. In mice, recall responses are studied within weeks, maximally months after the immunization; in humans, years or decades might have elapsed between the last antigen encounter of the T cell and its reactivation for the testing purpose.
Cytokine production of antigen-specific Th1/Th2/Th17 memory cells is best studied in freshly isolated PBMC and relies on techniques permitting analysis at the single cell level. Intracytoplasmic cytokine staining (ICS) in conjunction with flow cytometry is one frequently used approach to accomplish this goal [20
]. ICS has a detection limit of around 1 in 1000 events, that is 0.1% [20
]. Antigen-specific Th2 and Th17 cells, however, rarely reach frequencies of even 0.1% in PBMC, and therefore mostly go undetected by ICS. Better suited for the Th1/Th2/Th17 delineation of memory cells is ELISPOT [21
]. In this approach, PBMC are plated on a polyvinylidene fluoride (PVDF) membrane that has been pre-coated with cytokine-specific capture antibody. When the PBMC are present between 100,000 and one million cells per well, they form a monolayer on the membrane and contacts between T cells and antigen presenting cells (APC) are secured [22
]. Once antigen is added, the antigen-specific T cells become activated and start to secrete the cytokine that they were preprogrammed to express while the cytokine is captured around the secreting cell by the cytokine-specific capture antibody on the membrane. Thus, each cytokine secreting cell leaves behind on the membrane a “cytokine spot” that can be visualized via addition of a cytokine-specific detection antibody. Counting these “spots”, also called spot-forming units (SFU), permits detection of the individual antigen-specific T cells secreting a particular cytokine, and the SFU count per well establishes the frequency at which those antigen-specific T cells occur within all PBMC plated in that well. As every cytokine producing cell is visualized, with 400,000 PBMC plated per well (as was done in this particular study), the detection limit of the ELISPOT assay is 1 in 400,000 cells, that is, the ELISPOT assay as performed was 400 times more sensitive than ICS. For ICS, toxic secretion inhibitors need to be added to retain the cytokine in the cell, limiting the antigen stimulation period to 6–8 h. In ELISPOT assays, no such additional reagents are used, and subsequently the antigen challenge period can be readily expanded to several days. Relying on the sensitivity of the ELISPOT assay and extending the observation period after antigen challenge we set out to optimize the detection of rare antigen-specific Th2 and Th17 cells.
In the present study, we selected 12 common recall antigens. Five of them were proteins that need to be processed by APC before they are presented on MHC class II molecules to CD4 cells [23
]. These are dust mite antigen (DM), purified protein derivate of mycobacterium tuberculosis (PPD), ultraviolet light (UV)-inactivated, so called “grade 2” HCMV virions (CMV gr.2), gamma radiation inactivated mumps virions, and mosquito antigen. In addition, pools of 15 amino acid long peptides were used that cover the protein sequence in 11 amino acid overlaps. Such peptide pools were used for the following proteins: BZLF1 and EBNA1 (both open reading frames of the Epstein Barr Virus, EBV), MP1H3N2, H1N1 and NPH3N2 (all three are proteins of flu virus). Finally, we also tested a 15-mer peptide pool that covers the CMV pp65 antigen sequence. Due to their length, 15-mer peptides activate CD4 cells [24
]. After screening a library of healthy human donors to identify individuals who possess Th1/Th2 and Th17 memory cells specific for these antigens, we set out to study the kinetics of antigen-triggered IFN-γ, IL-2, IL-4, IL-5 and IL-17 production by the respective CD4 memory cell subset. The question was asked whether kinetics differ between individual donors and antigens. The data showed that, invariably, IL-4, IL-5 and IL-17 secretion by CD4 cells is delayed by several days compared to the production of IFN-γ. This notion has profound implications for reliable detection of Th2 and Th17 cells as these T cell subsets would go undetected in assays that do not extend an 8 h antigen stimulation period. Accounting for the delayed secretion kinetics of Th2 and Th17 cells, therefore, should be part of any immune monitoring approach.
Our data shows that Th1, Th2 and Th17 cells do not produce cytokine synchronously. The secretion of IFN-γ by Th1 cells peaks 24 h after antigen encounter. While peptide-triggered IFN-γ production by these cells can reach close to maximal levels at 6 h—the time point typically used for detecting this cytokine by ICS, it requires 24 h for protein antigen that—unlike peptides—needs to be internalized, processed and transported on HLA Class II molecules to the surface of the APC before the antigen becomes available for T cell recognition. IL-2 production by Th1 cells also peaks at 24 h. For Th2 cells, the kinetics of IL-4 and IL-5 production showed marked differences. IL-4 production was detectable already at 24 h, but peaked at 48 h, whereas IL-5 secretion was barely detectable even at 48 h, and peaked at 72 h, and in some instances even later. The IL-17 secretion kinetics by Th17 cells was similar to the one of IL-5. A striking observation was that these kinetics can have antigen-dependent variations. The likely explanation for this finding is that—unlike peptides—complex antigens acquired by APC from the extracellular space are not internalized, processed and presented at the same rate. The APCs’ antigen uptake by pinocytosis will be substantially less efficient and slower than internalization of an antigen that binds to receptors on these cells. The pathway of antigen uptake can also impact the lysosomal processing of antigen, whereby receptor-mediated uptake is likely to activate the APC and the processing machinery, whereas pinocytotic uptake will leave the APC in a resting state. APC activated by antigen will also upregulate the expression of their cell adhesion and costimulatory molecules, thereby enhancing antigen presentation—unactivated APC will not do so. Some complex antigens, as used here, might even actively inhibit antigen processing/presentation. This seems to be the case for dust mite antigen, which causes delayed T cell responses for all cytokines tested.
Considerable interindividual variations have been seen in the kinetics of the T cell recall responses, even to the same antigen. These are likely to result from allelic polymorphisms affecting key molecules involved in the antigen processing and presentation machinery, not T cell activation itself. The frequency of the relevant alleles in the test population would then define the numbers of donors that need to be tested to establish the variation in kinetics.
Overall, the data highlights the necessity of establishing for each complex antigen the kinetics of cytokine production byT cells before testing a larger cohort of human donors for the recall response to that antigen.