Restoration of the peace in autoimmune disease: tired of fighting?

Few topics in T-cell immunology have received as much attention in recent years as the phenomenon known as “exhaustion”. This condition is widely believed to result from excessive repetitive stimulation of T cells, which “exhausts” their ability to perform “effector functions”, such as the production of proinflammatory cytokines, and makes them ineffective in the battle against certain viruses or cancers [1]. Some consider the term “exhaustion” too anthropomorphic and prefer to call this phenomenon “dysfunction” [2]. A study by Bae et al. on this issue of Cellular and Molecular Immunology presented evidence that at least some CD4⁺ T cells with the characteristics of “exhaustion” are not actually dysfunctional but instead serve a different function, namely, to promote the restoration of immunological tolerance after a temporary breach thereof [3].

The immune system has the delicate task of protecting against pathogens and cancer without causing excessive damage. Healthy tissue must not be attacked, necessitating the existence of mechanisms to ensure self-tolerance. To this end, self-reactive T cells are eliminated during thymic development, whereas regulatory T cells (Tregs) inhibit untoward immune responses via self-reactive T cells that inadvertently escape thymic selection [4]. These mechanisms generally work well, but the existence of autoimmune diseases shows that they are not always perfect. Furthermore, protracted immune responses can cause cumulative tissue damage and, at some point, it may be more beneficial to the fitness of the host to accept the presence of an invader than to endure the consequences of endless battle. T-cell “exhaustion/dysfunction” is thought to help with this, as the loss of T-cell effector functions limits their capacity to inflict damage. Why then even keep such exhausted T cells around? One thought is that they still help contain persistent viruses to subclinical levels without reaching a sterile cure [1, 2]. Another possible reason is proposed by Bae et al. [3].

Even more than antigens from chronic viral infections, self-antigens resist elimination. If self-reactive T cells respond to such antigens, one might expect them to eventually also become exhausted. Exhaustion is essentially a passive tolerance mechanism: the cells merely abstain from doing damage. Bae et al. suggested that exhausted cells can also play an active role in restoring tolerance [3]. This discovery was made via a mouse model for multiple sclerosis (MS) known as experimental autoimmune encephalomyelitis (EAE), in which CNS inflammation is induced by immunization with the CNS antigen myelin oligodendrocyte glycoprotein (MOG). In the version of this model used in this study, mice develop a single episode of neurological symptoms (e.g., paralysis), followed by a spontaneous remission phase during which the immune response in the CNS subsides (Fig. 1). Bae et al. reported that self-reactive CD4⁺ T cells found in the CNS during this remission phase exhibit signs of exhaustion: they express the inhibitory receptors PD1, Tim3 and Lag3; possess a limited capacity to proliferate in vitro; and are unable to produce the proinflammatory cytokines IFNγ and IL-17 (both of which are known to contribute to the development of CNS pathology) [3, 5]. These cells are characterized by expression of the Notch2 surface receptor and by low expression of FoxP3, a transcription factor best known for its role in Tregs [4]. The cells identified by Bae et al. are not Tregs but nonetheless appear to play a tolerogenic role. The transfer of Notch2⁺ CD4⁺ T cells, isolated from the CNS of mice in the remission phase of EAE, reduces the development of pathology in MOG-immunized recipient mice [3]. Moreover, deletion of Notch2 in cells expressing FoxP3 leads to exacerbation of pathology in such mice [3]. Since FoxP3⁺ Tregs do not express Notch2 [3], these results are consistent with the interpretation that Notch2 regulates the ability of these Notch2⁺FoxP3^low CD4⁺ T cells to promote the restoration of immune tolerance.

The tolerogenic activity of Notch2⁺FoxP3^low CD4⁺ T cells may be explained at least partially by their ability to increase the number of Treg cells in the inflamed CNS. They appear to do so via several mechanisms (Fig. 1). First, Notch2⁺FoxP3^low CD4⁺ T cells suppress conventional effector T cells and induce their conversion into Tregs, presumably through surface expression of TGFβ [3]. Moreover, Notch2⁺FoxP3^low CD4⁺ T cells appear to increase the proliferation of Tregs via TNF-α, which is tethered to their plasma membrane [3]. While secreted TNF is proinflammatory and acts via the activation of TNFR1, membrane-bound TNF has a much greater affinity for a second receptor, TNFR2 [6]. This receptor is prominently expressed by Tregs, and its activation potently costimulates their expansion [7]. The role of membrane TNF-α in restoring tolerance likely has implications for the treatment of autoimmune disease [7]. TNF-α- antagonist antibodies are used to treat various inflammatory disorders, including rheumatic and inflammatory bowel diseases. Intriguingly, such treatment actually exacerbates the clinical symptoms of MS [8].

Tolerogenic Notch2⁺FoxP3^low CD4⁺ T cells are not unique to this EAE mouse model: phenotypically similar cells were found in the colons of mice and human patients with inflammatory bowel disease, suggesting that they may play a general role in downregulating autoimmune responses. Further support for this idea comes from an earlier study that demonstrated the presence of seemingly exhausted, self-reactive FoxP3^low CD4⁺ T cells in the blood of patients suffering from several other autoimmune diseases [9]. The expression of Notch2 was not reported in this latter study but may have been missed for technical reasons. As Bae et al. reported, Notch2 mRNA is equally expressed in Notch2⁺ and Notch2^– CD4⁺ T cells [3]. However, Notch2 protein is selectively preserved in Notch2⁺ cells because of the low abundance of the ubiquitin ligase Nedd4, which targets the Notch2 protein for degradation [3]. The preferential presence of Notch2 in these cells would therefore not be detected by transcriptome analysis, the main method used to characterize cell populations comprehensively.

How are these Notch2⁺FoxP3^low CD4⁺ T cells related to “traditional” exhausted CD4⁺ T cells, such as those found inside tumors and in the context of chronic viral infection [10, 11]? Bae et al. themselves are careful not to call them exhausted (they call them dysfunctional, which in my view also does not do them justice) [3], but striking similarities exist, as described above. On the other hand, some differences exist, among which the fact that Notch2⁺FoxP3^low CD4⁺ T cells lack the elevated expression of Tox family transcription factors that regulate the traditional exhaustion program [1, 3]. The late appearance of self-reactive Notch2⁺FoxP3^low cells in the inflamed CNS fits intuitively with the idea that these cells are descendants of effector T cells that first attack the tissue and eventually become exhausted. However, this idea is not supported by trajectory analysis on the basis of transcriptome composition, which suggests that Notch2⁺FoxP3^low cells follow a separate developmental path [3]. Although such analyses are certainly not definitive proof, it should be noted that “traditional” exhausted T cells are currently also seen as a separate lineage, with dedicated precursors that diverge from the regular effector cell differentiation path early in a response rather than developing from effector cells [1]. It is intriguing to speculate that Notch2⁺FoxP3^low cells belong to a novel lineage that exists to restore tolerance. Whether or not a separate lineage, it seems safe to predict that we will not have heard the last of the interesting cells discovered by Bae et al. [3]. In fact, given the parallels with exhaustion, much attention may be still to come.