Regulation unmasked by activation - Nature

6 downloads 0 Views 630KB Size Report
Mike Clark and Anne Cooke are in the. Department of Pathology, University of Cambridge,. Cambridge, UK. e-mail: [email protected]. Despite the widespread ...
ne w s an d v ie w s

Regulation unmasked by activation Mike Clark & Anne Cooke The glycoprotein CD52 is an important target for clinical antibodies, but its receptor and function have remained a mystery. However, it now seems that CD52 may be released in soluble form by a subpopulation of human T cells and may thereby exert an as-yet-unrecognized regulatory function via the inhibitory molecule Siglec-10.

npg

© 2013 Nature America, Inc. All rights reserved.

D

espite the widespread clinical use of antibodies that target the glycosylphosphatidylinositol (GPI)-anchored glycoprotein CD52, little is known about its function. In this issue of Nature Immunology, Bandala-Sanchez et al. provide evidence that the lectin Siglec-10 is a receptor for soluble human CD52 (ref. 1). Soluble CD52 is released in a phospholipase C–dependent manner from a population of CD4+ T cells with high expression of CD52; it then binds to Siglec-10 on activated T cells, which results in inhibition of their function. The Siglec family comprises sialic acid– binding proteins present mainly on cells of the innate immune system. Siglec-10 is expressed on several subsets of human leukocytes, including B cells, eosinophils, monocytes and some natural killer cells2, but published studies have not found Siglec-10 on T cells. However, the finding that Siglec expression may be ‘unmasked’ after removal of endogenous sialylated ligands when cells are activated suggests that Siglec-10 expression may have been overlooked, as published studies have used resting, not activated, T cells3,4. Bandala-Sanchez et al. now demonstrate Siglec-10 expression on activated human T cells but not on their resting counterparts1. In terms of structure, Siglec-10 has the potential to interact with sialic acid residues on CD52. An immunoreceptor tyrosine-based inhibitory motif present in the cytoplasmic tail of Siglec-10 inhibits signaling through its recruitment of the tyrosine phosphatases SHP-1 and SHP-2, as well as of the inhibitor of cytokine signaling SOCS3 (ref. 3). Structural and functional considerations therefore support the proposal that Siglec-10 acts as an inhibitory receptor. Siglec-10 has been shown to bind to the adhesion molecule VAP-1 and the cell surface marker CD24 (refs. 5,6). VAP-1 expression is upregulated during inflammation, and VAP-1 is expressed on endothelial cells; moreover, Siglec-10 expressed on B cells acts as a substrate that leads to release of hydrogen peroxide, which facilitates further recruitment of leukocytes to inflammatory sites. Mike Clark and Anne Cooke are in the Department of Pathology, University of Cambridge, Cambridge, UK. e-mail: [email protected]

696

The mouse homolog of Siglec-10 is Siglec-G, and studies of targeted deletion of Siglec-G in mice, coupled with in vitro studies of samples from both humans and mice, have identified an important role for the interaction of Siglec-10 or Siglec-G, respectively, with CD24 in controlling the response to damage-associated molecular patterns. Bandala-Sanchez et al. begin by demonstrating that a CD52hi T cell subset in the peripheral blood of normal humans can directly inhibit the antigen-specific activation of effector (CD52lo) T cells1 (Fig. 1). However, both the frequency and function of these cells seem to be impaired in peripheral blood lymphocyte samples from patients with autoimmune diabetes. The authors then go to great lengths to demonstrate that this newly identified regulatory population is distinct from classic thymus-derived regulatory T cells (Treg cells). The authors also present data obtained with an adoptive-transfer model of diabetes in mice in which depletion of CD52hi T cells from donor populations leads to accelerated development of diabetes. Although that is consistent with the idea that the CD52hi subpopulation exerts some regulatory effect, the data obtained with mice are still preliminary, as their CD52hi T cells are not characterized in any detail. For example, they could be conventional thymusderived Treg cells, in which case the effects of their depletion would be expected. Another outstanding question is how a putative CD52hi Treg cell would operate in mice. Would it necessarily act via the mouse Siglec-10 homolog Siglec-G? Further analysis of these cells in mice will doubtless shed light on such issues. CD52 is expressed on many cells of the immune system; for example, there is high CD52 expression on B cells and T cells and lower CD52 expression on monocytes, macro­ phages and eosinophils, with weaker CD52 expression on natural killer cells and sometimes CD52 expression on granulocytes but no CD52 expression on hematopoietic progenitors or stem cells7. It is thus interesting that the cell types with surface expression of CD52 substantially overlap those known to express Siglec-10, the proposed receptor for soluble CD52. Crosslinking of CD52 by antibody activates both CD4+ T cells and CD8+

T cells8. As noted above, CD52 is a GPI-linked glycoprotein; in its mature form, it is composed of 12 amino acids with one amino-linked glycosylation site through which a complex carbo­ hydrate terminating in sialic acid is bound9. Two distinct forms of CD52 have been isolated (CD52-I and CD52-II) that differ in the phosphatidylinositol moieties of their GPI anchor and their sensitivity to phosphatidylinositolspecific phopholipase C10. The data of BandalaSanchez et al. suggest that it is the isoform of CD52 that is sensitive to phosphatidylinositolspecific phospholipase C that has a key role in its interaction with Siglec-10 on T  cells, as immunoregulation in vitro is ablated by inhibition of phospholipase C activity1. To further examine the mechanism of suppression, the authors generate a soluble fusion protein of CD52 and the immunoglobulin crystallizable fragment (CD52-Fc) that binds to Siglec-10 on activated T cells and inhibits the T cell response to both foreign and self antigens, as well as proximal T cell signaling events such as phosphorylation of the tyrosine kinase Zap70. Furthermore, blockade of Siglec-10 with a specific antibody negates the regulatory effects of CD52-Fc, but blockade of several other human Siglec proteins (Siglec-5, Siglec-7, Siglec-9 and Siglec-14) does not, which suggests that the CD52 acts exclusively via Siglec-10. Although the data on the specificity of the CD52–Siglec-10 interaction are clear, many questions remain about its in vivo relevance. For example, are the amounts of CD52-Fc needed to mediate suppression in vitro (micrograms) really achievable physiologically? Indeed, concentrated conditioned medium from activated CD52hi T cells is needed to even detect the soluble form, and the authors report that such concentrated preparations are insufficient for use in suppression assays in vitro. Another intriguing question raised in the study by Bandala-Sanchez et al. is what exactly these CD52hi T cells are1. Although they seem to be distinct from ‘classic’ Treg cells, it is important to remember that definitions of ‘Treg cell– ness’ are not as conclusive in humans as they are in mice. Are these cells, then, something akin to an induced Treg cell—in other words, an essentially ‘standard’ effector cell that then ‘decides’ to adopt a regulatory fate on the basis

volume 14 number 7 JULY 2013 nature immunology

ne w s an d v ie w s ?

?

Siglec-10 CD52 CD52hi T cell

Inhibition

Phospholipase C

APC

CD52lo T cell

Soluble CD52

npg

© 2013 Nature America, Inc. All rights reserved.

Inhibition

Figure 1 Possible mechanisms of suppression by soluble CD52. CD4+ T cells are activated by antigen presented by antigen-presenting cells (APC) such as dendritic cells. Phospholipase C cleaves CD52 from CD52hi CD4+T cells. Soluble CD52 binds to Siglec-10 expressed on activated T cells and inhibits the T cell response. Given the high expression of Siglec-10 on cells of the innate immune system, it seems likely that CD52 may also act via such cells. Additionally, Siglec-10 expression on activated T cells raises the possibility that CD52 acts in an autocrine manner, although the present paper by Bandala-Sanchez et al.1 does not directly address either of those possibilities.

of CD52 release? If so, then what are the triggers for this ‘decision’? Are they ‘instructive’ or stochastic? If the CD52hi T cells are indeed a distinct regulatory cell population, then it would be important to delineate their roles and physiological importance relative to that of thymus-derived Treg cells and the various forms of induced Treg cells, both Foxp3– and Foxp3+. A further unanswered question is where the soluble CD52 acts physiologically. Although the in vitro data obtained by Bandala-Sanchez et al. demonstrate a direct effect of CD52-Fc and CD52hi T cells on effector T cells1, given the expression pattern of Siglec-10, we speculate that there is also an important component of suppression exerted by cells of the innate immune system. This seems likely, given the abundance of data in the literature showing that Treg cells do indeed modulate the function of dendritic cells; perhaps ligation by soluble CD52 is yet another mechanism by which this can occur. Alemtuzumab, the humanized monoclonal antibody to CD52, has therapeutic efficacy in a range of conditions, including chronic

lymphocytic leukemia and multiple sclerosis7. Generally, very high doses of alemtuzumab are used for the treatment of chronic lymphocytic leukemia, whereas, in contrast, much smaller doses have been used to treat multiple sclerosis in clinical trials. Such trials have shown that after the depletion of lymphocytes by antiCD52, the repopulating T cells are mainly of a memory or regulatory phenotype (that is, CD4+CD25hiFoxp3+). The present paper raises the interesting question of the susceptibility of the proposed CD52hi regulatory cells to lysis during in vivo treatment with alemtuzumab. The target of alemtuzumab is thought to include the GPI glycan anchor of CD52 and its carboxy-terminal tripeptide near the lipid bilayer of the cell membrane, with its aminolinked oligosaccharide not being needed for its antigenic activity. Bandala-Sanchez et al. suggest the intriguing possibility that removal of the CD52hi Treg cell population could be responsible for the autoimmune pathologies that arise in some patients with multiple sclerosis treated with alemtuzumab1. Whether this is indeed the case or whether it is the

nature immunology volume 14 number 7 JULY 2013

consequence of recovery of the immune system from lymphopenia remains to be clarified. Other unanswered questions include whether both isoforms of CD52 are expressed on all T cells and whether isoform expression changes after activation, whether expression of phosphatidylinositol-specific phospholipase C (and of which phospholipase C isoforms) is altered during T cell activation and whether there is the possibility of an autocrine effect. The contrasting roles of Siglec-10 in facilitating inflammation through its interaction with VAP-1 and inhibiting inflammation through its interactions with CD24 (coupled with damageassociated molecular patterns) and, as shown in this manuscript1, CD52, also need to be put into context. The identification of Siglec-10 as a receptor for soluble CD52 opens up exciting new areas of study about the mechanisms that control the phosphatidylinositol-specific phospholipase C–mediated cleavage of GPIlinked CD52, as well as finer details about the mechanism by which the response of target T cells is inhibited. Examination of the relative contributions of a direct effect of CD52 on Siglec-10 expressed on T cells versus an indirect effect of CD52 via dendritic cells opens up further avenues of exploration, not the least being characterization of the effects of soluble CD52 on dendritic cell function and that of other Siglec-10-expressing cells of the innate immune system. COMPETING FINANCIAL INTERESTS The authors declare competing financial interests: details are available in the online version of the paper (doi:10.1038/ni.2646). 1. Bandala-Sanchez, E. et al. Nat. Immunol. 14, 741–748 (2013). 2. Cao, H. & Crocker, P.R. Immunology 132, 18–26 (2010). 3. Razi, N. & Varki, A. Glycobiology 9, 1225–1234 (1999). 4. Crocker, P.R., Paulson, J.C. & Varki, A. Nat. Rev. Immunol. 7, 255–266 (2007). 5. Chen, G.Y., Tang, J., Zheng, P. & Liu, Y. Science 323, 1722–1725 (2009). 6. Kivi, E. et al. Blood 114, 5385–5392 (2009). 7. Coles, A.J. Neurotherapeutics 10, 29–33 (2013). 8. Rowan, W.C., Hale, G., Tite, J.P. & Brett, S.J. Int. Immunol. 7, 69–77 (1994). 9. Xia, M.Q. et al. Biochem. J. 293, 633–640 (1993). 10. Treumann, A., Lifely, M.R., Schneider, P. & Ferguson, M.A. J. Biol. Chem. 270, 6088–6099 (1995).

697