- Top of page
- Supporting Information
EAE is the primary pre-clinical disease for modelling the autoimmune/inflammatory component of multiple sclerosis. In fact, EAE is the primordial CD4+ T-cell-driven autoimmune disease model. It is striking (although perhaps unsurprising) that more than 10 000 publications over seven decades have provided a confusing, rather than a satisfying, picture of etiopathology. In the current issue of the European Journal of Immunology, an analysis of mice lacking LFA-1 is reported. Given the role of this integrin in T-cell activation and effector cell migration, one might predict resistance of LFA-1−/− mice to EAE induction. Instead, this study unexpectedly reports that EAE was exacerbated in the absence of LFA-1, and that this correlated with a decrease in the steady-state numbers of Foxp3+ Treg in the LFA-1−/− mice. Previous studies on the role of LFA-1 in EAE have been reviewed recently. This Commentary focuses on our current understanding of Treg function in the development and resolution of EAE and discusses how the absence of LFA-1 might unhinge these, possibly by altering the generation of Treg in the thymus, their expansion in response to autoantigen immunization, or their infiltration of the CNS.
Studies on protective lymphocyte populations in EAE can be traced back to the T suppressor days of the 1970s (reviewed in 1). The current consensus is that a variety of T-cell populations can either have natural regulatory functions or can be experimentally coerced into providing them. Amongst these populations, Foxp3+ Treg have primacy, in that they have a clear capacity to limit the initial expansion of encephalitogenic T-effector cells and are essential to the natural resolution of inflammation in monophasic EAE models, by functioning specifically within the CNS 1, 2. Myelin-responsive natural Treg (nTreg) appear to be constituents of the “steady state” nTreg pool and are able to expand in response to antigen in vivo3. Thus, the ultimate clinical outcome of immunizing a mouse with a myelin autoantigen reflects an “arms race” between the T-effector cells and Treg. Several lines of evidence indicate that even a transient impairment of Treg can allow the T-effector cells a head start, resulting in a more severe pathology that cannot be easily remedied, either by the Treg themselves or by other regulatory mechanisms.
In this issue of the European Journal of Immunology, Gültner et al. 4 describe a way in which the balance can be tipped in favour of the T-effector cells, leading to enhanced pathology. LFA-1, a heterodimer composed of CD11a (αL integrin) in combination with the common β2 integrin chain (CD18), plays important roles in T-cell activation. In addition to providing a key component of the immune synapse through the interaction with ICAM-1 on APC and other cells, LFA-1 signalling can potentiate TCR-driven activation by enhancing MAPKinase and AP-1 activity 5–7. LFA-1 also has roles in the migration of T cells (and other immune cells) into the tissues 8. Thus, there are several potential points at which the genetic deficiency in LFA-1 could abrogate the development of EAE. Contrary to our expectation, in the study by Gültner et al., 4 myelin oligodendrocyte glycoprotein (MOG)-induced EAE was shown to be exacerbated in mice deficient in the LFA-1-specific CD11a integrin, rather than diminished, ruling out an essential role for LFA-1 in T-effector activation and infiltration of the CNS. In fact, as is usually the case when more severe and/or prolonged clinical signs are evident, the LFA-1−/− mice had more inflammatory lesions containing greater numbers of other immune cells as well as CD4+ T cells within their CNS, and greater demyelination. Within the increase in overall CD4+ T-cell numbers in the CNS, there were also more MOG-responsive T cells. Thus, the enhanced disease seen in the absence of LFA-1 seems to reflect the burden of T-effector cell numbers, rather than an overt qualitative change in their function, as the levels of IFN-γ, IL-17, and TNF-α produced in response to MOG were unaltered in the KO mice.
Since our initial report on Foxp3+ Treg accumulation and function within the inflamed CNS during EAE 9, flow cytometric analysis of this population of cells has become a reasonably straightforward protocol. In their study, Gültner et al. 4 found that similar overall numbers of Treg entered the inflamed CNS irrespective of the LFA-1 status, but that the increased T-effector cell infiltrate led to a reduced frequency of Foxp3+ cells within the CNS CD4+ infiltrate of the LFA-1−/− mice.
In response to CNS inflammation, Treg migrate into the CNS and enter a rapid proliferative phase allowing them to achieve higher CNS frequencies than those seen in the peripheral lymphoid organs 9, 10. Thus, in passive EAE induced by adoptive transfer of MOG-responsive T-effector cells, as many as 50% of the CNS-infiltrating CD4+host cells can be Foxp3+10. Because of their rapid in situ proliferation, it has been difficult to determine how many Treg actually migrate from the blood, to establish the CNS Treg response. The homing molecules that are needed by Treg to enter the CNS are not fully defined, but the study of Gültner et al., 4 suggests that this does not require LFA-1. A recent report has suggested that CCR6 has a role in Treg entry into the CNS during EAE 11.
It has been our assumption that the rapid proliferation of Treg within the inflamed CNS is driven by – as yet, undefined – factors that are produced within the inflammatory lesions. Although it remains unclear whether antigen recognition is an essential component of this proliferative trigger, CNS Treg in EAE strongly express CD103, suggesting a recent TCR-triggering 9. This begs the question as to why Treg in LFA-1−/− mice fail to achieve the same frequency in the CNS as the one seen in WT mice? Do a finite number of Treg enter the CNS and these Treg cannot keep up with the greater load of infiltrating T-effector cells in the KO mice? Do LFA-1−/− Treg have an intrinsically deficient proliferative burst? Do Treg require LFA-1 for their suppressive function in the CNS and in the absence of LFA-1 the T-effector cells are unstoppable? All these questions are testable. The latter prospect, i.e. that LFA-1 has a direct role in the suppressive function of Treg, has been suggested by studies showing that the Treg-mediated down-modulation of CD80 and CD86 on splenic DC is dependent on not only CTLA-4, but also LFA-1 12.
Gültner et al. 4 demonstrate that not only were there many more MOG-responsive T-effector cells in the CNS of the LFA-1−/− mice, but this was also mirrored in the spleen and the lymph nodes. This observation strongly suggests that, in the absence of LFA-1, the T-effector cells gain an early head start during the initial few days after immunization. Could a Treg deficit at this point account for the exacerbated pathology in the KO mice? Gültner et al. 4 found an ∼50% reduction in the total number of CD4+Foxp3+ cells in the spleens of LFA-1−/− mice and a similar reduction in the frequency of the Foxp3+ cells amongst CD4+ thymocytes, indicating an impaired capacity of the KO mice to generate nTreg. This is consistent with a similar effect seen in mice lacking the common β2 integrin chain, CD18 13. Furthermore, CD18−/− Treg had impaired suppressive function in vitro and in vivo13 although, from that study, the defect could not conclusively be ascribed to a specific LFA-1 deficiency.
LFA-1 and other β2 integrin family members have been investigated extensively, with some conflicting results in EAE (recently reviewed in 14). As Gültner et al. 4 acknowledge, the reasons for the discrepancies between their data and those of others are not clear. An aside worth noting is that, in the study by Gultner et al. 4, the clinical disease in the WT mice was relatively mild, and the disease seen in their LFA-1−/− counterparts was more in line with what one might call “normal” EAE.
Similar studies from the Barnum group 15 have found clinical EAE to be lessened in LFA-1−/− mice, rather than worsened, after active immunization with MOG. However, exacerbated disease followed transfer of WT MOG-responsive T-effector cells into LFA-1−/− hosts, which is consistent with the Gultner study 4. This exacerbated disease could well be attributable to ineffective Treg function in the hosts, since bioluminescent imaging revealed rapid activation of the transferred cells in the lymphoid system followed by their markedly enhanced infiltration of the CNS 15. The Barnum group have also found reduced Treg numbers in the spleen (but not the thymus) of LFA-1−/− mice, together with impaired in vitro suppressive function of LFA−/− Treg 16. The observed activation of the transferred effector cells within the lymphoid organs prior to their colonization of the CNS may seem strange. In fact, this was first described some time ago in rats by the Wekerle group 17, although the functional significance of this observation remains unclear.
So, LFA-1−/− mice have fewer Treg and show increased severity of EAE under a sub-optimal induction protocol. Partial depletion of Treg using an anti-CD25 antibody rendered WT mice susceptible to EAE that was similar to (in fact, more rapid than) the disease seen in LFA-1−/− mice 4. Although this is not a formal proof that defective Treg function underlies the enhanced pathology in the KO mice, it is at least supportive of this possibility. Many studies have reported that anti-CD25 treatment exacerbates EAE 9, 18, 19. Perhaps most relevant to the observations of Gültner et al., 4, we previously found that anti-CD25 treatment allows development of clinical EAE in mice receiving a sub-agonist variant of a myelin basic protein peptide. In Treg-sufficient mice, this variant does not provoke disease 20. Treg are therefore most probably able to limit the clonal expansion of T-effector cells to sub-optimal TCR stimulation, consistent with their influence on DC activity, as discussed above.
The possibility that LFA-1 deficiency might impair the ability to generate nTreg in the thymus is intriguing. A contemporary paradigm is that Foxp3 expression is triggered in those developing thymocytes that receive TCR signals just below the threshold for negative selection, thus diverting the most self-reactive T cells that survive into the peripheral nTreg pool 21. Presumably, this process is responsible for the generation of the MOG-responsive nTreg that we know exist within the peripheral T-cell repertoire of C57BL/6 mice 3. If LFA-1 signalling potentiates TCR signalling within thymocytes, then, in the absence of LFA-1, those thymocytes that would normally receive a TCR signal of sufficient strength to trigger Foxp3 expression might instead populate the naïve (non-Treg) peripheral T-cell pool. This might account for the enhanced pathology seen in the Gültner study 4 in two ways – fewer MOG-responsive nTreg and an expanded precursor pool for the T-effector response to MOG immunization in the absence of LFA-1. There is a precedent for thymocyte expression of surface molecules, other than TCR, influencing the generation of myelin-responsive Treg. The Wraith group 22 recently reported that myelin basic protein-responsive TCR transgenic mice produce a ∼5-fold greater frequency of CD4+Foxp3+ thymocytes when crossed onto a CTLA-4−/− background. This reverse situation in the absence of a key co-inhibitory molecule would be consistent with the possibility outlined above. Our own work and other studies 3 have found that 20–30% of the Foxp3+ cells found in the CNS during active EAE induced with MOG are responsive to MOG and that these are nTreg that expand in response to MOG immunization. Whether these MOG-responsive Treg are present or impaired in LFA-1−/− mice would be an important extension of the Gültner study. To conclude, the Gültner study 4, together with other recent reports, suggests that LFA-1 may have a key role in Treg function within the lymphoid system, in both active and passive EAE. Although an additional impact on the function of Treg within the CNS cannot be excluded, it seems that, once loose, the runaway T-effector train can become unstoppable. Perhaps of greatest importance, this study 4 underlines the need for new drugs that would allow the Treg to “catch up and apply the brakes”.