Adaptive immunity against pathogens originates with vigorous clonal expansion of antigen-specific T lymphocytes that recognize their cognate antigens on the surface of DC in secondary lymphoid organs. Once primed, CD4+ T cells migrate to follicles to help B cells produce antibodies, and to peripheral sites of antigen exposure to fight incoming pathogens by delivering the appropriate type of effector cells function. Type 1 effector Th1 cells produce IFN-γ that promotes clearance of viruses and intracellular bacteria, while Th2 cells produce IL-4, IL-5 and IL-13 that promote clearance of extracellular parasites 1, 2. Once antigen is eliminated, central memory (TCM) and effector memory (TEM) T cells persist in the memory pool to provide systemic immune surveillance in lymphoid organs and in peripheral non-lymphoid tissues to react promptly in case of re-attack 3. To do this T cells not only keep ‘memory’ of the cytokines to be produced, but also for the site where effector function is needed. This is the case for memory T cells that migrate to the skin or the gut 4.
The distinction of Th1 and Th2 cells and of TCM and TEM subsets provided the initial paradigms to understand the functional organization of the immune response 1, 3. Mouse and human studies have together contributed to the advancement of the field. The finding of Th1 and Th2 cells in mice 5 was soon matched with the description of similar subtypes of human T cells 6, while reciprocally the original description of TCM and TEM cells in humans 7 was confirmed and extended in the mouse model 8, 9.
Recent studies indicate that the binary paradigms of Th1 and Th2 cells and of TCM and TEM cells represent an over-simplification and point to the existence of multiple pathways of effector T-cell differentiation and multiple layers of memory T cells that provide tailored mechanisms of protection and immune surveillance against different pathogens in different tissues.
Expanding the Th1-Th2 paradigm: Th17, Treg and more
Based on seminal work on Th1 and Th2 cells, a T-cell lineage is defined as a cell population in which a change in cytokine production is promoted by polarizing signals and stably imprinted by a lineage-specifying transcription factor through epigenetic mechanisms 10. Th1 differentiation is promoted by IL-12 (and, in humans, also by type I IFN 11), and requires the transcription factor T-bet 12 that mediates inheritable modifications of the IFN-γ gene leading to its expression following antigenic stimulation. In contrast, Th2 differentiation is promoted by IL-4 and requires the transcription factor GATA-3 13 that mediates inheritable modifications of the IL-4, IL-5 and IL-13 genes 14.
Recently, two additional lineages have been described: Th17 cells, that produce IL-17 and IL-22, and antigen-induced Treg (iTreg). Mouse and human ThH17 differentiation are both dependent from the transcription factor ROR-γt 15, 16, but there is yet no clear consensus on the cytokines that promote their differentiation. Initial studies clearly identified IL-6 and TGF-β as Th17-promoting cytokines in mice 17–19, while in vitro experiments showed that IL-1β, IL-6 and IL-23 are required for differentiation of human Th17 cells 20, 21. Subsequent studies revealed an important role for IL-1β for mouse Th17 differentiation 22 and a requirement for low doses of TGF-β for human Th17 differentiation 23, 24, while high doses of TGF-β were shown to be inhibitory in both humans and mice 20, 25. The finding that circulating resting T cells in humans show constitutive TGF-β signaling 26 is puzzling and raises the issue of whether TGF-β is a rate limiting factor for human Th17 responses. Differences in the requirements for exogenous TGF-β between mice and humans may apply also to iTreg. Indeed, while TGF-β in the absence of IL-6 readily induces iTreg differentiation in mice 16, 18, there is no evidence that TGF-β may do the same in humans 27. The iTreg are functionally similar to thymus derived natural Treg in that they are anergic, suppressive, and capable of inhibiting disease in vivo. Differentiation of iTreg is under the control of the transcription factor FoxP3 in both mice and humans 16, and in mice can be triggered by continuous admnistration of low doses of peptide 28. The capacity to induce human antigen-specific Treg in vitro remains a long-sought goal for immunotherapy of autoimmune diseases.
There is increasing evidence that help to B cells is mediated by a dedicated lineage of follicular Th cells (TFH) that produce IL-21 and express ICOS and PD-1 29. These cells are characterized by the expression of CXCR5 that mediates their homing and long-term residence in the B-cell follicles 30–32. Recent studies suggest that IL-21 and Bcl6 may be involved in promoting and specifying differentiation of TFH cells 33, 34. Germinal center TFH cells appear to be only distantly related to circulating CXCR5+ T cells, as revealed by large scale gene expression analysis 35.
With the advent of multiparameter flow cytometry, it has become clear that individual cells can produce effector cytokines in different combinations 36, a fact that raises the question of whether there is heterogeneity within a lineage or whether each distinct cytokine-combination represent a separate lineage. For example, Th1, Th2 or Th17 cells that secrete IL-10 in a transient manner 37–39 may represent an altered state of T-cell activation rather than distinct cell fates, which is consistent with the finding that there is no cytokine memory for IL-10 expression 40. However, T cells stably producing high amounts of IL-10 and endowed with regulatory function can be induced in vitro and in vivo by antigenic stimulation in the presence of IL-10 or immunosuppressive drugs 41, 42. The finding that these cells do not express Foxp3, and suppress via secretion of IL-10 and TGF-β, has been taken as evidence that they represent a separate T-cell lineage (defined as Tr1) 43.
IL-22 was originally described in mice and humans as a cytokine characteristic of fully differentiated Th17 cells 39, 44. It was therefore surprising to find that a distinct subset of human skin-homing memory T cells produced IL-22, but neither IL-17 nor IFN-γ 45, 46. Differentiation of IL-22 producing T cells could be promoted by stimulation of naïve T cells in the presence of IL-6 and TNF or by plasmacytoid DC and appears to be independent from RORγt, but dependent on the aryl hydrocarbon receptor 45, 46. It remains to be established whether these cells, which have been operationally defined Th22, represent a lineage distinct from Th17.
Muddying the waters: Commitment and flexibility in cytokine gene expression
Ever since the discovery of Th1 and Th2 cells, it has been clear that there are human T-cell clones that simultaneously produce IL-4 and IFN-γ (defined as Th0). There are now further examples of T cells with a dual personality, even within memory T cells ex vivo. For instance, T cells producing both IL-17 and IFN-γ have been identified as a distinct subset of T cells in peripheral blood and in inflamed tissues 44, 47. Recent studies suggest that T cells can even change personality. This is the case for some Foxp3+ Treg that become Th17 cells 48, 49.
The presence of T cells that produce cytokines characteristic of two different lineages raises questions about the mechanisms that lead to their generation in vivo since Th1, Th2 and Th17 fates appear to be mutually exclusive. A possible explanation is that T cells committed to a certain lineage maintain the memory of the originally imprinted cytokine, but at the same time have the capacity to acquire expression of additional cytokines if stimulated under appropriate polarizing conditions. Thus, when cloned under neutral conditions, Th1 or Th2 cells retain IFN-γ or IL-4 producing capacity, respectively; however, when stimulated under polarizing conditions opposite to their phenotype (i.e. in the presence of IL-4 for Th1 and in the presence of IL-12 for Th2) Th1 acquire the capacity to produce IL-4 in addition to IFN-γ and, reciprocally most Th2 cells acquire the capacity to produce IFN-γ together with IL-4 50. Other examples of flexibility include Th17 cells that, when stimulated with IL-12, become IL-17/IFN-γ dual producer cells 47.
In summary, it appears that flexibility is part of the differentiation program so that initially T cells maintain open differentiation options, while at later stages they may become irreversibly committed to one lineage. Considering that in the human system substantial numbers of memory T cells are uncommitted and that most cells maintain cytokine flexibility, it is possible that expression of opposing cytokines may be induced at different times, in different tissues or enforced by delivering recall antigens together with appropriate polarizing signals.
Further heterogeneity revealed by homing receptors
The function of effector and memory T cells is dependent not only on the expression of cytokines, but also on the capacity to migrate to sites of antigen encounter. T-cell migration is dependent on the expression of a particular set of selectins, chemokine receptors and integrins that determine, in a combinatorial fashion, the steps of extravasation and positioning in different tissue microenvironments 51. Numerous studies have addressed in past years how homing receptors are selectively expressed and regulated in different subsets of T cells 3, 52.
Chemokine receptors have been particularly useful for dissecting T-cell subsets with distinct migratory capacity and effector function. These studies, mostly performed in the human system, are the basis for our current distinction between TCM, TEM and TFH and for our capacity to discriminate between Th1, Th2 and Th17 cells and gut- and skin-homing T cells using surface markers. CCR7 and CD62L have been used to discriminate between TCM and TEM3. Human and mouse studies demonstrated that TCM are involved in recall responses in secondary lymphoid organs having high proliferative and reconstituting capacity, while TEM are present primarily in peripheral tissues and are endowed with immediate effector function, but show poor reconstituting capacity. CXCR3, CCR4 and CXCR5 identify further subsets within TCM; CXCR3+ and CCR4+ TCM do not produce effector cytokines, but spontaneously differentiate to Th1 and Th2, respectively, while CXCR5+ TCM represent uncommitted non-polarized memory T cells 53. Within TEM CXCR3 and CCR5 define primarily Th1 cells, whereas some T cells expressing CCR3, CCR4 or CRTh2 are Th2 54.
The study of the antigenic repertoire of phenotypically distinct T-cell subsets has provided relevant insights into the human immune response. This approach led to the original observation that memory T cells specific for mycobacterial antigens are Th1, while T cells specific for extracellular parasites and allergens are Th2 6. Since then, this approach has been used to analyze both protective and pathogenic T-cell responses. Recently it has been shown that memory T cells specific for Candida albicans are present almost exclusively within the CCR6+ CCR4+ T-cell subset that produces IL-17, while M. tuberculosis purified protein derivative-specific memory T cells are present within a subset of CCR6+ CXCR3+ T cells that produce primarily IFN-γ 44. In contrast, human cytomegalovirus-specific CD4+ T cells are present with the pre-Th1 and Th1 subsets while tetanus toxoid-specific T cells are found in virtually all circulating memory T-cell subsets, including CCR7+ TCM cells, CXCR5+ TFH and CCR6+, as well as CCR6−, TEM cells 44. It can be envisaged that this approach will continue to provide relevant information of the human immune response in infections, vaccinations and diseases.
The differential expression of adhesion and chemokine receptors on T cells and the constitutive expression of the corresponding ligands in different tissues allows the discrimination of T cells that provide immune surveillance in different organs 4. The integrin α4β7 and CCR9 are expressed by T cells that migrate to the gut, where the corresponding ligands MadCAM and CCL25 are expressed by intestinal endothelial and epithelial cells. Conversely, CCR4, CCR10 and CLA are expressed on T cells that migrate to the skin. Homing of memory T cells to the skin provides an example of how different receptors can be used sequentially in the process of extravasation and tisssue positioning. Entry of circulating T cells is mediated by the adhesion molecules CLA interacting with vascular E-selectin, and requires also interaction of α4β1 with VCAM, and of LFA-1 with ICAM-1. CCR4 55 and potentially CCR8 56 are required for the transition from the blood to the dermis; whereas, CCR10 is required for targeting T cells from the dermis to the epidermis where its ligand CCL27 is produced by keratinocytes 57. Remarkably, the gut and skin-homing receptors CCR9 and CCR10 are induced in T cells by vitamin A and vitamin D, which are selectively produced in the gut and skin, respectively 4. Recent studies suggest that CCR6+ T cells may provide immune surveillance in the brain by crossing the choroid plexus epithelial cell barrier that constitutively expresses the CCR6 ligand CCL20 58, while a similar function in the liver may be carried out by CXCR6+ NKT cells that patrol liver sinusoids that express CXCL16, the cell surface ligand for CXCR6 59.
In summary, there is little doubt today that the regulation of homing receptor expression is an integral part of the T-cell differentiation program, although the exact mechanisms and the precision of this regulation remain to be established.
Functional modules of T-cell immunity
To provide a framework to understand the heterogeneity of CD4+ T cells we propose an organization based on distinct functional modules (Fig. 1). Each T-cell module is tailored for a particular class of immune response that provides either regulatory function (e.g. B-cell help or suppression) or effector function against different types of pathogens. The function of each module requires a combination of molecular and cellular interactions that can be described in terms of polarizing cues (usually a cytokine or cytokine-combination), lineage-specifying transcription factor, homing receptors and effector molecules (usually cytokines) and finally cross-talk with other immune or tissue cells.
The value of a module is the demonstration that each of its components is essential for a particular type of immune response. These components have been extensively validated for Th1 and Th2 cells, which represent modules involved in responses to intracellular microbes and extracellular parasites, respectively. A more recent addition to these modules, namely Th17, represents a functional module required for immunity to fungi. This view is supported by the fact that fungi provide a selective stimulus for Th17 differentiation 44, that memory T cells specific for C. albicans are primarily Th17 and that patients lacking STAT3 do not have Th17 cells and are selectively susceptible to severe fungal infections 60–62.
While for some modules there is a general consensus on all or most of the above aspects, for others some pieces of information are still missing. Compelling evidence for a TFH module comes from experiments where the lack of CXCR5 or SAP was shown to impair antibody responses by preventing migration into B-cell follicles and sustained T-cell–B-cell interaction 63–65. However, there is yet no clear consensus on the nature of the polarizing cues and of the lineage determining transcription factor. Similarly, while there is compelling evidence for a role of the Treg modules in regulating immune response to self and foreign antigen, there are still uncertainties as to the nature of the polarizing stimuli and the suppressor effector mechanisms.
Finally, and in spite of still fragmentary information, it is tempting to speculate that Th22 cells may represent a module responsible for skin health and homeostasis. This view is supported by the selective expression of skin-homing receptors combined with the production of a cytokine, IL-22, that stimulates keratinocyte motility and production of antimicrobial peptides 66. The fact that the immune system appears to dedicate a functional module for skin health highlights the broad control that the immune system exerts over bodily function. Interestingly, the Th22 module has a counterpart in NK-22 cells, an NK-cell subset that has recently been described in the gut of mice and humans and that it is thought to be involved in the regulation of intestinal epithelial cell function 67–69. Thus, Th22 and NK-22 cells may represent two functionally overlapping modules of adaptive and innate immunity, respectively.