See accompanying article: http://dx.doi.org/10.1002/eji.201242824
CD28 costimulation: Walking the immunological tightrope
Article first published online: 16 JAN 2013
© 2012 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
European Journal of Immunology
Volume 43, Issue 1, pages 42–45, January 2013
How to Cite
Sansom, D. M. and Walker, L. S. K. (2013), CD28 costimulation: Walking the immunological tightrope. Eur. J. Immunol., 43: 42–45. doi: 10.1002/eji.201243211
- Issue published online: 16 JAN 2013
- Article first published online: 16 JAN 2013
- Manuscript Accepted: 4 DEC 2012
- Manuscript Received: 28 NOV 2012
- Manuscript Revised: 28 NOV 2012
- Costimulatory molecules;
- Regulatory T cells;
- T cells
CD28 function has typically been associated with the generation of effector T-cell responses to Ag. However, it is also clear that CD28 plays an important role in Treg-cell biology. Understanding which functions predominate is important when designing therapeutic interventions based on CD28 targeting. An article by Hünig and colleagues [Eur. J. Immunol. 2013. 43: 188–193] in this issue of the European Journal of Immunology uses an inducible gene deletion approach to reveal that, in the steady state, Treg cells intrinsically require CD28 signals for their maintenance in the periphery, whereas homeostasis of conventional T cells is relatively unaffected. Here we highlight the delicate balance created by the ability of CD28 to modulate both regulatory and effector T-cell responses.
The CD28/CTLA-4 pathway represents the ultimate immunological balancing act. While CD28 is required for effective T-cell responses to pathogens, CTLA-4 (CD152) is essential to prevent T-cell activation, most notably to our own bodies. One might imagine that we perturb this balance at our peril, yet this pathway represents one of the most targeted areas in therapeutic immunology. A number of important drugs have emerged based on this pathway including the immune suppressive Abatacept , which blocks CD80 and CD86 and is in use for the treatment of rheumatic diseases , as well as the immune-activating (CTLA-4-blocking) antibodies such as Ipilimumab, which are being used in cancer immunotherapy [3, 4]. Given the extensive targeting, it is perhaps surprising that, as yet, we still do not have a comprehensive picture of how the four main interacting components (CD80, CD86, CD28, and CTLA-4) function and the precise consequences of these interactions. Moreover, the ill-fated CD28 agonist Ab TGN-1412 serves as a continuing reminder of the need for a complete understanding of this pathway . In this issue of the European Journal of Immunology, Hünig and colleagues  add further information to this important puzzle providing additional detail on the impact of CD28 on Treg-cell homeostasis.
At its most basic, the CD28 pathway provides a simple solution to a critical problem confronting the immune system: whether or not to activate a T-cell response? The problem arises since the generation of the T-cell repertoire in the thymus inevitably leads to selection of self-reactive T cells, which must ultimately be controlled when they enter the peripheral circulation . One solution to this problem is to constrain T-cell activation by requiring additional costimulatory signals to reinforce the signals from TCR engagement that together then lead to T-cell activation. These costimulatory signals are provided by CD28 interacting with its ligands CD86 and CD80 on APCs . Since elevated levels of CD28 ligands are wired into “danger signals," such as TLRs or inflammatory cytokines sensed by APCs, this effectively provides confirmation of the need to trigger a robust effector T-cell response.
Despite its intuitive appeal, the situation above is over-simplified and several additional complexities are apparent. Firstly, CD28 also shares its two ligands with a high-affinity homologous receptor (CTLA-4) that is an essential inhibitor of T-cell responses. The importance of CTLA-4 is amply demonstrated by the observation that CTLA-4-deficient mice die by 3–4 weeks of age of profound autoimmunity [9, 10]. How CTLA-4 prevents autoimmunity has been the subject of considerable debate ; however, what is clear is that perturbing ligands can affect both CD28 and CTLA-4 function. A second issue is that CD28 and CTLA-4 are both implicated in Treg-cell homeostasis and function. It has been previously shown that mice that constitutively lack either ligands or CD28 have fewer Treg cells [12, 13] due to effects in both the thymus and the periphery [13, 14]. Finally, the impact of CD28 and CTLA-4 on thymic selection in general still remains to be fully elucidated. Thus interfering with the CD28/CTLA-4 pathway potentially perturbs both Treg and T effector cells influencing their generation, peripheral homeostasis, and ultimately function. Unraveling these various issues has, therefore, presented a considerable challenge.
In their paper, Hünig and colleagues  provide some further insights into the role of CD28. By using a tamoxifen-inducible deletion system, the authors studied the impact of CD28 deletion after an initial cohort of T cells had passed through the thymus with CD28 intact. The authors reveal that deletion of CD28 has a significant impact on the numbers of Treg cells, but not on conventional CD4+ or CD8+ T cells, in the periphery; the decrease becoming more profound over time following tamoxifen treatment. This confirms that CD28 plays an important role in Treg-cell homeostasis. Moreover, by performing thymectomy, the authors eliminated the possibility that deletion of CD28 in the thymus was responsible for the loss of Treg cells in the periphery, again emphasizing the postthymic impact of CD28 on Treg-cell numbers. More significantly, by making bone marrow (BM) chimeras where CD28 was lacking in only ∼50% of the cells, the authors could also demonstrate that the Treg-cell requirement for CD28 was clearly cell-intrinsic and could not be rescued by the presence of CD28-expressing cells. Consistent with this, we also observe cell-intrinsic defects in Treg cells in mixed BM chimeras using CD28-deficient and sufficient cells (Walker unpublished observation). Together these data reinforce the concept that postthymic cell-intrinsic CD28 signals in Treg cells are necessary for the maintenance of Treg-cell numbers in the periphery (Fig. 1).
Hünig and colleagues  also identified phenotypic differences in CD28-deficient Treg cells, as these cells expressed lower levels of CD25 and CTLA-4 in comparison with their CD28-sufficient counterparts. While reduced expression of CD25 could potentially be responsible for the impaired peripheral homeostasis of CD28-deficient Treg cells, the BM chimera experiments suggest otherwise. In fact, expression of CD25 was restored to normal levels when CD28-deficient Treg cells existed in the presence of WT T cells. This result likely reflects a role for IL-2 produced from WT (but not CD28-deficient) T cells in maintaining Treg-cell CD25 levels . However, IL-2 derived from WT cells is, nonetheless, unable to compensate for CD28 costimulation in maintaining peripheral Treg-cell numbers. This situation mirrors previous observations by Singer and colleagues , who demonstrated impaired thymic development of CD28−/− Treg cells that could also not be corrected by coexistent WT cells. Thus it appears that there are nonredundant, cell-intrinsic roles for CD28 in Treg-cell generation and homeostasis, as well as cell-extrinsic effects of IL-2 in their maintenance.
Given that CD28 is clearly required for Treg-cell homeostasis, one might also wonder about the impact of CTLA-4 on this process. There are already several insights here from the literature. When CTLA-4-deficient mice were first reported they were characterized, in addition to their early lethality, by an obvious expansion of CD4+, CD25+ T cells [9, 10]. Quite reasonably, this was thought to relate to an expansion of effector T cells involved in auto-immune pathology. While, in part, this is true, more recent evidence using Foxp3 staining indicates that many of these expanded cells are in fact Treg cells [17, 18]. This observation fits very well with the concepts raised by Hünig and colleagues , in that CD28 is required to promote Treg-cell proliferation and survival, a process presumably driven by the shared ligands which bind to both CD28 and CTLA-4. Given that CTLA-4 is the higher affinity receptor constitutively expressed by Treg cells, its removal presumably results in an increase in CD28 signaling, thereby accounting for the Treg-cell expansion observed in CTLA-4-deficient mice. In addition, the fact that CD28 costimulation can augment Treg-cell proliferation also suggests that Treg cells are constantly being triggered by some form of TCR engagement. Indeed it is known that Treg cells proliferate in response to self-Ags [19, 20] and the evidence now suggests this proliferation is costimulated by CD28. Thus, rather ironically, considering their anergic reputation, it seems that in the steady state Treg cells turn over more dynamically in vivo than conventional T cells, utilizing both TCR and CD28 signals.
In addition to effects on Treg-cell homeostasis, Hünig and colleagues  also report that Treg cells conditionally deleted for CD28 show impaired suppressive function. One explanation for this impaired suppression is that it relates to the lower levels of CTLA-4 expression observed. Multiple reports have linked CTLA-4 expression to Treg-cell suppressive capacity [21-23] and it seems that, despite being expanded in the context of CTLA-4 deficiency , these Treg cells exhibit impaired function. The reduced levels of CTLA-4 expression seen by Hünig and colleagues  could result from a number of possible mechanisms. In conventional T cells, transcription of CTLA-4 is linked closely with T-cell activation and proliferation, it is therefore possible that CTLA-4 transcription in Treg cells is also CD28-dependent. In addition, activation signals are also known to influence CTLA-4 posttranslationally since intracellular trafficking of CTLA-4 protein is linked to T-cell activation [24, 25].
How CTLA-4 acts as an effector molecule for Treg-cell function has also been an active area of research [11, 26]. Recently, we identified that CTLA-4 is capable of transendocytosis, a process whereby it acts like a “molecular hover” physically removing CD80 and CD86 from APCs . According to this model, Treg cells can dominantly regulate T-cell responses by ensuring that, in the steady state, there is insufficient ligand available to trigger CD28 costimulation. Thus, one major control point for both immunity and autoimmunity is the level of CD80 and CD86 ligands in the system. Interestingly, it would appear that in the process of regulating the availability of CD80 and CD86 molecules, Treg cells can simultaneously receive CD28 signals themselves that contribute to their own homeostasis. Importantly, CTLA-4 expressed by Treg cells appears to attenuate, rather than completely block, CD28 signals as indicated by enhanced proliferation following CTLA-4 blockade or deletion [17, 22]. Treg cells are, therefore, presumably capable of utilizing lower levels of CD28 signaling, as their own CTLA-4 impairs CD28-ligand interactions. This ability of Treg cells to utilize reduced CD28 signals is compatible with the concept that they express TCRs with higher affinity for self-Ags than their conventional T-cell counterparts . Overall, there appear to be considerable similarities between conventional T cells and Treg cells regarding the use of CD28; however, whereas conventional T cells rely on CD28 to drive expansion in the context of foreign Ags, Treg cells use this pathway to maintain their numbers in response to self-Ags.
Taken together, the above discussion highlights the complexity in selectively manipulating any single aspect of CD28/CTLA-4 function. For example, based on the observations of Hünig and colleagues , one would predict that treatment with drugs such as abatacept or belatacept would ultimately result in a loss CD28 signals and diminished Treg-cell numbers. However, it is necessary to understand that any loss of regulation is counterbalanced by the fact that effector T cells are also functionally impaired, with the net result that immune responses are suppressed. This explains the general lack of autoimmunity in mice lacking CD86 and CD80, despite an obvious deficit of Treg cells in these animals . Equally, treatment with anti-CTLA-4 would be expected to enhance Treg-cell numbers by promoting CD28 function, yet this is counterbalanced by blockade of their CTLA-4-dependent suppressive function, justifying anti-CTLA-4 as an immune enhancing drug for tumor therapy. Thus, as our knowledge of the subtleties of the CD28/CTLA-4 pathway increases, along with our knowledge of its impact on multiple aspects of the T-cell immune response, it will hopefully be possible to more precisely tailor interventions to achieve the desired therapeutic outcomes.
LSKW is an MRC senior research fellow.
Conflict of interest
The authors declare no financial or commercial conflict of interest.
- 5The storm has cleared: lessons from the CD28 superagonist TGN1412 trial. Nat. Rev. Immun. 2012. 12: 317–318.,