Special regulatory T-cell review: T-cell dependent suppression revisited

Abstract

The concept of T-cell dependent regulation of immune responses has been a central tenet of immunological thinking since the delineation of the two cell system in the 1960s. Indeed T-cell dependent suppression was discovered before MHC restriction. When reviewing the data from the original wave of suppression, it is intriguing to reflect not just on the decline and fall of suppressor T cells in the 1980s, but on their equally dramatic return to respectability over the past decade. Hopefully their resurgence will be supported by solid mechanistic data that will underpin their central place in our current and future understanding of the immune system.

Cannon to right of them,

Cannon to left of them,

Cannon in front of them

Volley’d and thunder’d

Storm’d at with shot and shell,

Boldly they rode and well,

Into the jaws of Death,

Into the mouth of Hell,

Rode the six hundred (suppressionists).

(Adapted from The Charge of the Light Brigade, Alfred, Lord Tennyson)

Phase 1: Nascency (1970–86)

T-cell-dependent suppression had its origins in the wake of the discovery of the two-cell system when it became apparent not only that T and B lymphocytes talked to each other,^1,2 but that self-tolerance could no longer simply be explained in Burnetian terms of repertoire purging during early life – interestingly an idea to emanate from the striking association of autoimmunity with immunodeficiency states in humans.³

At the first Brook Lodge Symposium on Immunological Tolerance held in 1974 after publication of the original papers by Gershon and Kondo⁴ and McCullagh,⁵ the entire international community of suppressionists could be accommodated in one modestly sized room. However, by the time of the 1976 Cold Spring Harbor Symposium, there was standing room only and the podium was shared between us and the major histocompatibility complex (MHC) restrictionists.

The initial focus of work on suppression was directed primarily to models of tolerance in the dual contexts of T–B collaboration and the Bretscher and Cohn⁶ two-signal hypothesis. In our laboratory, two experimental systems were used, one involving a soluble foreign antigen (human gamma globulin) and the other a self antigen (mouse red cells).

Model 1: Human immunoglobulin G

Human immunoglobulin G (HGG) in deaggregated form induced T-cell-dependent suppression⁷ in addition to tolerance in both T-cell and B-cell compartments.⁸ The main characteristics of the phenomenon are summarized in Table 1. Essentially, suppression was antigen-specific and was mediated by T cells expressing Ly-2 (CD8) (suppressor T cells or Ts cells). An apparent lack of classical MHC restriction was explained by the presence of antigen (or anti-idiotype) on the Ts-cell surface, leading us to propose a series of possible mechanisms involving cell contact and/or release of a soluble factor (Fig. 1a).⁹ Importantly, a population of long-lived Ly-2⁻ quiescent memory Ts cells expressing Ly-1 (CD5) was identified in vivo in tolerant hosts.^10–12 These memory Ts cells were thought to be responsible for the accelerated suppression observed on re-exposure to antigen and in fact the tolerant state correlated more closely with memory than effector Ts cells¹² (Fig. 1b). Primary and secondary suppression could be induced by antigen not just in deaggregated (tolerogenic) form, but when aggregated,¹³ passaged transplacentally to mimic self antigen¹² or given orally.¹⁴ Further refinement of the importance of antigen presentation came from experiments by Sercarz’s group demonstrating the apparent existence of suppressor versus helper epitopes on well-defined proteins such as hen egg lysozyme and β-galactosidase,^15,16 a phenomenon that remains of interest to this day, along with that of ‘epitope-specific regulation’, which emphasizes the negative outcome of excessive carrier priming, e.g. in vaccination protocols (reviewed in ref. 17).

Table 1.

T suppressor (Ts) cells versus T helper (Th) cells¹

Characteristics	Ts	Th
Phenotype	Th-1+, Ly-1−/2,3+	Th-1+, Ly-1+/2,3−
Effector	L3T4−, I-J+	L3T4 +, I-J−
Memory	as above but Ly-1+	as above
Tissue distribution	Spleen (mainly),gastrointestinaltract2	Spleen, lymphnode, thoracicduct, blood
Antigen specificity	Yes	Yes
Target	B cell, but requiredTh to suppress	B cell
MHC restriction	No	Yes
Relative radiosensitivity	Sensitive	Resistant
Colchicine sensitivity	Yes	No
MLR reactivity	No	Yes
Concanavalin Areactivity	Yes	Yes
Suicide with 131Iantibody	Yes	No
Cell surface adherence	Yes	No

¹See references 1,7,9–14,23.

²Site of CD8αα T cells (IELS).

MHC, major histocompatibility complex; MLR, mixed lymphocyte reaction.

Figure 1.

(a) Schematic diagram illustrating three possible models for the mechanism of suppression in human immunoglobulin G (HGG) tolerance. Ts = suppressor T cell; Th = helper T cell; anti-Id = anti-idiotype; Id = idiotype. (Reproduced with permission of the authors and publishers from Ref. 9, p. 98.) (b) Schematic diagram of tolerance and suppression after prenatal administration of dHGG. Percentages refer to duration of tolerance which is dose dependent in contrast to suppression. (Reproduced with permission of the authors and publishers from Ref. 12, p. 234.)

Thus a clear-cut role had been established for Ts cells (particularly memory Ts cells) in maintaining tolerance and regulating collaborative antibody responses. Other investigators independently reached a similar overall conclusion using a range of protein antigens and also extended the concept of suppression to anti-idiotype¹⁸ and anti-allotype responses¹⁹ as well as Ir gene-linked non-responsiveness.²⁰ Interestingly, immunoglobulin E (IgE) production (and by inference allergic reactions) was shown early on to be peculiarly susceptible to suppression.^21,22 Consequently Taylor and one of us (A.B.) were prepared to make the following statement in a review article published in 1976:²³

‘Suppression seems to be a component of most types of immune response whether tending towards immunity or tolerance’.

A question often asked is why did most reports at that time describe Ts cells as expressing CD8 rather than CD4? The experimental models were generally designed to measure foreign antigen-specific induced Ts cells rather than ‘natural’ suppressors with self-specificity. Recent reports have confirmed that a variety of induced regulatory T-cell subsets do express CD8 (see Phase IV below). A second factor was the ease with which CD8-expressing T cells could be depleted by cytotoxic antibodies to the allo-determinants Ly-2 (CD8α) and Ly-3 (CD8β), whereas murine CD4 is monomorphic, preventing the generation of allo-specific antibodies. Instead, the allelic marker Ly-1 (CD5) was used as a positive marker of CD8⁻ T cells. However difficulties arose from the fact that it was also expressed, albeit at a lower level, by CD8⁺ cells. In addition, when the first rat anti-mouse CD4 antibody became available, it did not work reliably for cell depletion, particularly in vivo.

Model 2: Autoimmune haemolytic anaemia (AIHA)

The second system we used was the AIHA model²⁴ in which T-dependent regulation of a true anti-self response to mouse red blood cells was demonstrable following immunization with cross-reactive rat red cells. In this case the precise phenotype of the Ts cells was less clear cut because distinctions were starting to be drawn between ‘suppressor inducers’ and Ts effectors, although the consensus at the time favoured expression of Ly-2 by the latter.

Despite the controversy that arose later concerning I-J, polyclonal anti-I-J serum (raised by immunizing B10.A3R recipients with cells from B10.A5R donors for anti-I-J^K and vice versa for anti-I-J^b) was remarkably effective, both in vivo and in vitro, in reversing suppression and breaking tolerance, particularly when given perinatally in the AIHA model.^25,26 The significant findings to emerge from our studies in this model are summarized in Table 2 and clearly demonstrated the importance of T-cell-dependent suppression in true anti-self responses.

Table 2.

Anti-self T suppressor (Ts) cells

Characteristics	References
Memory Ts cells for self antigens exist innormal hosts	9
Depletion of Ts cells during ontogeny but toa lesser extent in adult life leads to a loss ofself tolerance, manifest by spontaneousdevelopment of autoantibodies	Reviewed in26
A higher proportion of B-cell hybridomasgenerated in absence of memory Ts cells secreteantibodies with anti-self specificity
Transfer of cross-reactive T-cell clones selectivelyreduces anti-self (murine red blood cells) butnot the anti-foreign (rat red blood cells)antibody response, an effect transferable withsplenic T cells to secondary recipients.

Clinical implications

Based on these studies and those of other investigators, we were in a position as early as 1980 to publish a review²⁷ drawing attention to the relevance of T-cell-dependent suppression in a range of disease states including: immunodeficiency states, aplastic anaemia, chronic infections and malignancy, in addition to autoimmune and allergic diseases and transplantation tolerance. The spectrum of diseases on this list is remarkably similar to those linked to regulatory T (Treg) cell abnormalities in recent times, although the limited availability of markers for human Ts cells retarded progress in the 1970 and 1980s.

Phase II: Decline and fall (1982–1988)

The fact that T-cell-dependent suppression had become a robust holistic concept at the epicentre of immunology made its subsequent decline and fall all the more intriguing. The reasons largely revolved around the failure of Ts cells to comply with the requirements of the new molecular immunology. No gene(s) encoding I-J could be identified within the MHC,^28,29 messenger RNA for the β-chain of the T-cell receptor (TCR) was not present in Ts hybridomas,³⁰ no unique markers for Ts cells had been discovered, and attempts to define Ts-cell function in vitro led to the development of implausibly complex regulatory circuits dependent on poorly defined I-J⁺ soluble factors. These issues are well summarized in an issue of the Scandinavian Journal of Immunology featuring the Cartesian onslaught led by Göran Möller³¹ and responses from some of the more temperate members of the international immunological community.^32–34 Ironically the non-believers were just as guilty as the believers in assuming the existence of a discrete and homogeneous suppressor subset, which with hindsight we know to be a gross oversimplification. With hindsight, one might speculate that the I-J region contains non-coding DNA sequences involved in regulating the expression of the molecules recognized by the anti-I-J antisera.

Phase III: Interregnum: all was not lost (1985–1995)

Fortunately the concept of immune regulation was never fully extinguished. Thus in the 1980s and early 1990s, immune deviation (reviewed in ref. 35) became explicable in terms of the T helper type 1 (Th1)/Th2 paradigm,^36,37 while convincing evidence of a role for CD4⁺ T cells in specific allograft tolerance as well as rejection was provided by groups in the transplantation field^38,39 (reviewed in ref. 40). The demonstration of CD25-expressing, self-reactive CD4⁺ Ts cells by Sakaguchi’s group in mice⁴¹ and of self-reactive Ts cells by Mason’s group in rats⁴² firmly re-established the relevance of T-cell-dependent regulation in self-tolerance and autoimmunity. The result: the reinstatement of immune regulation in ‘respectable’ circles with the re-emergence of Ts cells, now ‘rebranded’ as Treg cells, reminiscent of Gershon’s original suggestion to call them ‘regulator’ cells. The immunological community owes a real debt to these noble warriors who defended the pass at the Thermopylae of suppression so gallantly during the onslaught of the molecular non-believers.

Phase IV: Treg redux: has real progress been made? (1996–)

The major difference between now and then is the creation of a much more sophisticated framework within which to conduct definitive experiments involving well-defined cytokines, dendritic cell–T-cell interactions, transgenic models and high-tech flow cytometry.

The immediate impact has been impressive. Immune regulation has been confirmed as an in vivo phenomenon of major import in immunopathic diseases as well as in transplantation tolerance. Moreover, the existence of naïve Treg cells as a discrete lineage with an anti-self bias has been established,⁴³ as has a major role for FoxP3 in transmitting inhibitory messages.^44,45 Interestingly B cells with their lower expression of CD80/CD86 appear to be more efficient antigen-presenting cells than dendritic cells for inducing effector Treg cells,⁴⁶ which could explain our previous observation that the B cell was the target of suppression in a hapten-carrier system involving bidirectional cognate interactions between Ts cells and B cells.

The challenge now is to delineate more precisely the part played by naive, effector and memory Treg cells in immune homeostasis and immunopathic disorders.

In immune homeostasis, recent studies of dendritic cell–CD4⁺ T-cell interactions and TCR signal strength have shed new light on the broad biological significance of immunoregulation. Thus Treg cells appear to play a crucial role in controlling the threshold for T-cell activation via dendritic cell costimulation (S-Y. Tan, H. Bolton, M. Mouawad, B. Roediger, C. Power, W.-P. Koh, H. Yagita, E. Shklovskaya, B. Fazekas de St Groth, unpublished data).⁴⁷ When Treg-cell activity is high, as is postulated to occur in response to infection in the developing world (B. Fazekas de St Groth, unpublished data), dendritic cell expression of costimulatory molecules is low (like that on B cells) and production of potentially harmful but relatively low-affinity self-reactive effector T cells is inhibited. Conversely, as occurs in the developed world where Treg activity is lower, T cells are more readily activated by dendritic cells expressing high levels of costimulatory molecules in addition to otherwise harmless antigens, like self-determinants and allergens, resulting in the generation of Th1 and Th2 effector T cells, respectively (Fig. 2a,b). This way of thinking is attractive because it provides a rational explanation for the inverse relationship between the incidence of autoimmune and allergic diseases on the one hand and infectious diseases on the other in the developed world (ref. ⁴⁸ and B. Fazekas de St Groth, unpublished data).

Figure 2.

(a) A model for T-cell receptor (TCR)-dependent regulation via dendritic cell–T-cell interactions. Regulatory T (Treg) cells may function by means of a negative feedback loop controlling expression of CD80/CD86 on the dendritic cell surface, by analogy with positive feedback providing helper function via CD40 ligand–CD40 interactions with the dendritic cells. (b) A model illustrating how Treg-mediated control of CD80/CD86 expression may control the threshold of antigen recognition, crucial for preventing the activation of low-avidity self-reactive T cells that are below the cut-off imposed during thymic selection (B. Fazekas de St Groth, unpublished data). Treg cells stimulated during high-affinity responses to microbes would increase the threshold (indicated by the green line) by reducing dendritic cell expression of costimulatory molecules. Conversely, in the absence of strong Treg-cell activity, the threshold of self-antigen recognition may drop below the thymic cut-off (indicated by the black line), allowing activation of low avidity anti-self T cells.

In immunopathic disease, one of the key issues that was never resolved in the previous era of suppression was whether changes in Ts-cell numbers and/or function were of primary pathogenic importance or simply secondary to the disease process itself. Patients with the primary immunodeficiency states, immune dysfunction, polyendocrinopathy, enteropathy, X-linked syndrome (IPEX),⁴⁹ autoimmune polyendocrinopathy–candidiasis–ectodermal dystrophy (APECED)⁵⁰ and Wiskott–Aldrich syndrome (WAS)⁵¹ provide convincing evidence of the serious consequences of a selective as well as a generalized defect in Treg cells. Variable results have been obtained in a range of other diseases because of the limited availability of reliable markers for differentiating human Treg cells from their activated CD4⁺ CD25⁺ counterparts. However, the recent availability of anti-FoxP3 antibodies and the finding that both naïve and effector Treg cells express relatively low levels of CD127 has allowed more accurate identification and isolation of Treg cells.^52,53 Evidence is now emerging that primary deficiencies in Treg-cell output are implicated in young patients with multiple sclerosis⁵⁴ and inflammatory bowel disease (N. Seddiki, S-Y. Tan, A. Smialkowski, M. Gourter, S. Dervish, W. Selby, M. Solomon, S. Lee, P. McKenzie, B. Fazekas de St Groth, unpublished data). Conversely, data are accumulating to link a relative increase in tumour-infiltrating Treg cells with a poor prognosis.⁵⁵ Moreover, the use of cytotoxic drugs and/or antibodies for reducing Treg activity in tumours is yielding promising results in experimental models and early human trials.^56,57

What are the risks?

Despite impressive progress, it is remarkable how little attention is paid to the old literature. As Aristotle remarked 2400 years ago: ‘those who ignore history are doomed to repeat it.’ Just as the era of suppression foundered on the complexity of ill-defined suppressor factors and circuitry, so the era of regulation risks becoming immersed in an ever expanding morass of effector Treg-cell subsets. Indeed what is the definition of a Treg cell now? In addition to CD4⁺ CD25⁺ FoxP3⁺ cells, Th3 and Tr1, we have double-negative Treg cells⁵⁸ and bona fide CD8αβ⁺ Treg cells in mouse and man, only some of which express FoxP3, as well as CD8αα⁺ cells recognizing non-classical MHC molecules⁵⁹ (perhaps our original non-MHC-restricted Ts cells), ICOS-inducible CD8⁺ PD1⁺ cells responsible for tolerance in a cardiac transplant model⁶⁰ and CD8⁺ Treg cells directed to both self^61,62 and foreign⁶³ antigens other than MHC. In a sense, most of the latter subsets are more sophisticated versions of the CD8⁺ Ts cells described in the 1970s and 1980s, while plasmacytoid dendritic cells have replaced what we used to call suppressor macrophages.⁶⁴ Nevertheless, these recent reports of novel Treg subsets do re-emphasize the fact that effector T cells can, depending on their cytokine profile and location, function in an inhibitory or stimulatory fashion. Among the more intriguing in vitro examples of dual function are: (1) T follicular helper cells that help B cells, but suppress conventional effector T cells in germinal centres;⁶⁵ (2) FoxP3⁺ Treg cells that not only induce Th17 aggressors but can, given the appropriate cytokine stimuli, themselves differentiate into Th17 producers.⁶⁶ Thus Gershon⁶⁷ got it right again when he coined the term ‘hermaphrocyte’ to describe such promiscuity.

In terms of immunological history, now as then, there are two schools of thought about Treg, née Ts, cells: the believers who are leading the charge of the ‘Heavy Brigade’ in support of Treg cells as unique and specialized immune effector cells, and the non-believers who suggest that they simply represent various manifestations of the immune system’s need to keep the brakes on potentially harmful responses. Whichever camp one is in, all should agree that the challenge is to capitalize on our improved understanding of regulatory mechanisms to develop new preventive and therapeutic strategies in transplantation, tumour immunology, vaccination, autoimmunity and allergy.