A type 2 cytokine axis for thymus emigration

White et al. describe a new mechanism of thymus emigration that is controlled by expression of the type 2 IL-4 receptor by thymic stroma and production of IL-4 and IL-13 by thymic-resident invariant NKT cells.


IntroductIon
Thymic organization and the availability of distinct cortical and medullary intrathymic microenvironments provide a specialized framework that guides developing thymocytes through multiple stages of migration, proliferation, and differentiation (Takahama, 2006;Boehm, 2008). Importantly, understanding mechanisms that control intrathymic T cell development requires identification of stromal cellexpressed regulators that mediate specific developmental events. For example, restricted expression of DLL4, CD83, β5t, and CXCL12 to the cortex (Plotkin et al., 2003;Murata et al., 2007;Hozumi et al., 2008;Koch et al., 2008;Liu et al., 2016;von Rohrscheidt et al., 2016) enables this site to mediate CD4 − CD8 − (double-negative [DN]) T cell commitment, preTCR-mediated maturation and positive selection of CD4 + CD8 + double positive (DP) thymocytes. Similarly, expression of Aire, costimulatory molecules and CCL19 and CCL21 (Degermann et al., 1994;Anderson et al., 2002;Ueno et al., 2004) in the medulla creates a site for tolerance induction and postselection development and migration (Cowan et al., 2013;Webb et al., 2016;Xing et al., 2016). Thus, correct positioning of immature DP thymocytes in the cortex and mature single-positive (SP) thymocytes in the medulla regulates intrathymic T cell development.
Few known factors control functional specialization of thymic microenvironments. Consequently, differing roles of stromal cells in thymocyte development are poorly understood, and thus, the identification of novel regulators of thymic stroma is essential in understanding thymic control of T cell production. Here, we show that the cytokine receptor component IL-4Rα is expressed in the thymus medulla, including a subset of medullary thymic epithelial cell (TEC [mTEC]), where it forms part of a functionally active type-2 IL-4R complex. Analysis of T cell development in Il4ra −/− mice revealed defects in thymus emigration that map to expression of IL-4Rα by the thymic microenvironment. We provide evidence that IL-4Rα influences thymic egress via a mechanism distinct from the S1P-S1P 1 axis and identify CD1d-restricted invariant NKT (iNKT) cells as key regulators of emigration by providing IL-4 and IL-13 to trigger type-2 IL-4R signaling. Collectively, type-2 cytokines from innate T cells are a novel component of mechanisms controlling αβT cell egress from the thymus. results and dIscussIon thymus medullary disorganization in Il4ra −/− mice To identify new regulators of thymus function, we analyzed tissue organization and thymocyte distribution in thymic sections from mutant mice in which thymocyte-stromal cross talk may be disrupted. Mice lacking IL-4Rα (Il4ra −/− mice) had disorganization of the thymic medulla, which contained epithelial-free areas lacking ERTR5 + mTEC (Fig. 1). Interestingly, these areas were not acellular cysts but contained In the thymus, stromal microenvironments support a developmental program that generates mature t cells ready for thymic exit. the cellular and molecular specialization within thymic stromal cells that enables their regulation of specific stages of thymocyte development is poorly understood. Here, we show the thymic microenvironment expresses the type 2 Il-4r complex and is functionally responsive to its known ligands, Il-4 and Il-13. absence of Il-4rα limits thymocyte emigration, leading to an intrathymic accumulation of mature thymocytes within medullary perivascular spaces and reduced numbers of recent thymic emigrants. thymus transplantation shows this requirement maps to Il-4rα expression by stromal cells, and we provide evidence that it regulates thymic exit via a process distinct from s1P-mediated migration. Finally, we reveal a cellular mechanism by which Il-4 + Il-13 + invariant nKt cells are necessary for Il-4rα signaling that regulates thymic exit. collectively, we define a new axis for thymic emigration involving stimulation of the thymic microenvironment via type 2 cytokines from innate t cells.
mature SP4 and SP8 thymocytes ( Fig. 1 A), including SP4 Foxp3 + Tregs (not depicted). Quantitative analysis showed individual Il4ra −/− thymic sections contained ∼8 mTEC-free areas with a mean size of 0.15 mm 2 , contributing to 20% of the total medulla area (Fig. 1 B). Approximately 75% of these areas were located within 100 µm of the corticomedullary junction (CMJ; Fig. 1 B), a site of thymic egress (Weinreich and Hogquist, 2008;Maeda et al., 2014). Despite these abnormalities, cortex-medulla separation remained intact, as did typical localization of DP and SP thymocytes in the cortex and medulla areas ( Fig. 1 A). Although medulla disorganization can be caused by altered mTEC development (Boehm et al., 2003), cortical TEC (cTEC) and mTEC lo /mTEC hi subsets were comparable in WT and Il4ra −/− mice (Fig. S1). Thus, absence of IL-4Rα causes alterations in the medullary distribution of SP thymocytes that are not explained by altered TEC development.
As increased SP thymocytes can be caused by altered thymic egress (Boehm et al., 2003), we looked for perturba-tion of this process. In Rag2GFP mice, GFP levels indicate thymocyte medullary dwell time and discriminate developing thymocytes from recirculating mature GFP − T cells. Moreover, unlike intrathymic FITC injection, GFP directly identifies recent thymic emigrants in a noninvasive manner and so avoids possible confounding side effects caused by surgical intervention (Boursalian et al., 2004;McCaughtry et al., 2007;Hauri-Hohl et al., 2014;Cowan et al., 2016). In WT Rag2GFP and Il4ra −/− Rag2GFP mice, the frequency of GFP − SP4 thymic cells was comparable (Fig. 2 G), indicating increased mature cells in the Il4ra −/− thymus is not due to enhanced peripheral T cell recirculation. Interestingly, GFP levels in mature CD62L + HSA − SP4 T-convs were significantly lower in Il4ra −/− mice compared with WT mice (Fig. 2 H). Moreover, although numbers of splenic SP4 T cells in Il4ra −/− mice were unaltered (Fig. 2 I), we saw a significant reduction in GFP + SP4 recent thymic emigrants (RTEs). These cells also had significantly lower GFP levels compared with WT, indicating a more-mature status (Fig. 2 J). Collectively, these data suggest that, in Il4ra −/− mice, prolonged medullary occupancy of mature SP4 T-conv leads to their intrathymic accumulation, which is reflected by reduced efficacy of thymic egress and diminished RTE numbers.
As thymic exit involves transit through the perivascular space (PVS) that surrounds blood vessels (Mori et al., 2007;Zachariah and Cyster, 2010), we further examined SP thymocyte accumulations in tissue sections from Il4ra −/− mice. Thymocyte accumulations were detected around CD31 + blood vessels, in between ERTR7 + basement membrane layers, suggesting that medullary abnormalities in Il4ra −/− mice are caused by increased thymocyte accumulations within enlarged PVS (Fig. 3 B). To directly assess that, we i.v. injected anti-CD4 antibodies to label cells in the thymic PVS (Zachariah and Cyster, 2010;Mouri et al., 2014). Il4ra −/− mice contained significantly higher proportions and numbers of SP4 thymocytes labeled with the injected antibody ( Fig. 3 C). Thus, the increase in SP thymocytes in Il4ra −/− mice is accompanied by their increased accumulation within thymic PVS, indicating a role for IL-4Rα in thymic egress.
To see whether defects in thymus emigration in Il4ra −/− mice map to a requirement for IL-4Rα expression by the thymic microenvironment, we transplanted 2dGuo-treated FTOC from Il4ra −/− mice into nude mice. Here, IL-4Rα expression by host-derived hematopoietic and nonthymic stromal cells remains intact. Importantly, analysis of grafts showed that Il4ra −/− thymuses contained intrathymic accumulations of SP thymocytes within mTEC-free areas (Fig. 4 D) and had increased mature CD62L + HSA − SP4 cells (Fig. 4 E). Together, expression of Il4ra, Il13ra1, and Il13ra2 by TEC, and their responsiveness to exogenous IL-4/IL-13 stimulation in vitro, provides evidence for their expression of a functional type-2 IL-4R complex. In addition, data from Il4ra −/− thymus transplant experiments indicates that normal thymic emigration requires type-2 IL-4R expression by the thymic microenvironment.
inKt cells produce type 2 cytokines for thymus emigration To examine the mechanism by which type-2 IL-4R signaling influences thymic egress, we used IL-13GFP reporter mice (Neill et al., 2010). We identified a small subset of CD45 + IL-13GFP + cells in the adult thymus at steady state ( Fig. 5 A). Phenotypic analysis showed most of these cells were mCD1d-PBS57 + iNKT cells (Fig. 5 A;Liu et al., 2006;Rossjohn et al., 2012;White et al., 2014). Importantly, we also detected expression of both Il4 and Il13 mRNA in mCD1d-PBS57 + iNKT cells (Fig. 5 B) and a dominant IL-4 + IL-13 + iNKT cell population after in vitro stimulation (Fig. 5 C). Interestingly, although earlier studies indicate that IL-4 production by thymic iNKT cells occurs in the steady state , culturing thymocytes with Brefeldin A, followed by flow cytometric analysis, did not reveal IL-13 protein expression in iNKT cells (not depicted). Thus, although our data indicate that thymic iNKT cells express Il13 mRNA and so may be primed to produce IL-13, their steady-state in vivo production of IL-13 protein may require their controlled stimulation by additional unknown interactions in the medulla that are absent from thymocyte suspensions. To directly assess the requirement for iNKT cells in thymic egress, we analyzed CD1d −/− mice (Mendiratta et al., 1997). Similar to Il4ra −/− mice, we saw mTEC-free areas containing SP thymocytes, accompanied by a significant increase in the most mature CD62L + HSA − SP4 T-conv thymocytes (Fig. 5 D). Together with the findings of others on intrathymic IL-4 production , these findings suggest that iNKT cells represent a cellular source of IL-4Rα ligands and that these cells are required to promote thymic egress of αβT cells.
Intrathymic CD1d-restricted iNKT cells are heterogeneous and include long-term resident cells of unknown function that can be identified as persisting, donor-derived iNKT cells within lymphoid thymus transplants (Berzins et al., 2006). To examine the possible relevance of these cells to thymic egress, we transplanted lymphoid CD45.1 + thymus lobes into CD45.2 + mice and, after 6 wk, identified long-term, thymic-resident iNKT cells as CD45.1 + mCD1d-PBS57 + cells (Fig. 5 E). Interestingly, and in agreement with studies identifying the potent cytokine secretion ability of long-term, thymic-resident iNKT cells (Berzins et al., 2006), most CD45.1 + iNKT cells remaining within the thymus grafts produced IL-4 and IL-13 after stimulation (Fig. 5 E), suggesting that thymic-resident iNKT cells are an intrathymic source of cytokine ligands for the type-2 IL-4R.
By investigating the specialization of thymic stroma, we show that TECs express the type-2 IL-4R, which is functionally active in response to its ligands IL-4 and IL-13. That Il4ra −/− mice display multiple and specific medullary defects, including the intrathymic accumulation of mature CD4 + thymocytes and diminished numbers of RTE, points strongly toward a role for the type-2 IL-4R in controlling thymic exit. Importantly, although our findings show that IL-4Rα is expressed by TECs, additional non-TEC stroma may also express the type-2 IL-4R and contribute to regulation of thymic output. Whether its role described here is exclusive to TECs or not, the identification of IL-4Rα as a new regulator of thymus emigration reveals a novel mechanism by which thymic stroma influences thymocyte development and represents an important step in understanding late-stage thymus function. Relevant to that, although experiments performed here used BALB/c mice that are biased toward type-2 cytokine production, we also saw SP thymocyte accumulations in and Il4ra −/− mice i.v. injected with anti-CD4PE (top). Quantitation of anti-CD4PE-labeled SP4 cells in WT and Il4ra −/− mice (bottom); n = 6 from two separate experiments. Surface levels of S1P 1 (D) and CD69 (E) in conventional SP4 thymocytes from WT and Il4ra −/− mice. (D) Gray histogram is peripheral blood CD4 + T cells; data represent at least three experiments, n = 6. (E) Gray histogram is CD69 on total SP4 thymocytes, Data represent at least three experiments, n = 10 for WT and Il4ra −/− mice. (F) Effects of FTY720 in WT and Il4ra −/− mice. (Top) CD4/CD8 in total thymocytes. (Bottom) CD62L/HSA in conventional SP4 thymocytes. (G) Proportions of total SP4 and CD62L + HSA + SP4 T-conv. Data represent at least two experiments, n = 8. Error bars indicate SEM. A Mann-Whitney nonparametric U test was performed. *, P < 0.05; **, P < 0.01; ***, P < 0.001.
How the type-2 IL-4R controls thymic emigration of SP thymocytes is unclear. However, comparable levels of S1P 1 and CD69 in WT and IL-4ra −/− SP4 thymocytes and successful blockade of emigration in Il4ra −/− mice by FTY720 argue that S1P-mediated emigration remains active. In addition, although blockade of S1P-mediated migration promotes T-lymphopenia (Zamora-Pineda et al., 2016), Il4ra −/− deficiency does not. Although the reasons for this difference are not clear, it perhaps indicates the influences of IL-4Rα and the S1P-S1P 1 axis on thymic emigration are distinct. Interestingly, given that RTE in Il4ra −/− mice show increased maturity as indicated by diminished Rag-2GFP levels compared with WT counterparts, it may be that type-2 IL-4R signaling controls the length of time mature thymocytes spend within the medulla to ensure a correct program of postselection maturation takes place, which may then determine the rate and efficacy of emigration. Interestingly, we found that IL-4/IL-13 stimulation of TEC induced expression of the chemokines CCL21 and CXCL10. Although CCL21 is a ligand for CCR7, a known regulator of thymic exit (Ueno et al., 2002), CXCL10 is a ligand for CXCR3, which controls intrathymic retention of iNKT cells (Drennan et al., 2009). These findings raise the possibility that type-2 IL-4R signaling in the thymus influences egress via multiple chemotactic mechanisms. First, it may control the production of chemokines that directly regulate the emigration of mature CCR7 + SP thymocytes. Second, as part of a positive feedback loop, it may boost TEC production of CXCL10, which acts to retain CXCR3 + iNKT cells in the thymus, where they can continue to trigger type-2 IL-4R signaling. Finally, thymocytes accumulating within the thymic PVS may also suggest that chemokines and/or other migratory factors controlled by IL-4Rα signaling are required for entry into the circulation.
That mTEC lo express the type-2 IL-4R may also be significant in explaining its role in thymic egress. Indeed, mTEC lo express CCL21 (Lkhagvasuren et al., 2013), which we show here can be induced by IL-4/IL-13, and its receptor (CCR7) is expressed by SP thymocytes and RTE (Ueno et al., 2002;Cowan et al., 2013Cowan et al., , 2016. Interestingly, LTβR also regulates both thymocyte emigration and CCL21 + mTEC lo (Boehm et al., 2003;Lkhagvasuren et al., 2013) and, as with Il4ra −/− mice, defective thymus egress in Ltbr −/− mice does not cause T lymphopenia (Boehm et al., 2003). However, notable differences exist between Ltbr −/− and Il4ra −/− mice. For example, although reduced mTEC may explain altered thymic egress in Ltbr −/− mice (Boehm et al., 2003;Lkhagvasuren et al., 2013), that cannot be the case for Il4ra −/− mice, in which TEC development is normal. Altered mTEC development and gross medulla disorganization in Ltbr −/− might also explain why intrathymic accumulations of thymocytes are more obvious in thymic sections of Il4ra −/− mice, where mTEC development and gross medullary architecture are normal. Finally, although expression of LTβR ligands maps to conventional SP thymocytes (Boehm et al., 2003;White et al., 2010), the cytokines that trigger type-2 IL-4R signaling for thymic egress are produced by iNKT cells, including long-term, thymic-resident cells. Thus, the type-2 IL-4R represents an important regulator of thymic microenvironments that enables innate T cells to regulate the thymic egress of conventional αβT cells.

MaterIals and MetHods Mice
The following mouse strains on a BALB/c background were used at 8-12 wk of age: Cd1d −/− (Mendiratta et al., 1997), Il4ra −/− (Mohrs et al., 1999), IL-13GFP (Neill et al., 2010). WT BALB/c littermates were used as controls. To examine thymocyte egress, we crossed IL-4Ra −/− mice to Rag2GFP mice (Yu et al., 1999). CD45.2 + C57BL/6, and congenic CD45.1 + BoyJ mice were used for thymus transplant experiments. For the generation of embryos, the day of vaginal plug detection was designated day 0. C57BL/6 IL-13 GFP/GFP mice were also used for analysis of thymus sections. Husbandry, housing, and experimental methods involving mice were performed at the Biomedical Services Unit at the University of Birmingham, in accordance with local ethical review panel and national Home Office regulations.

Ftoc
Embryonic thymus lobes at E15 (d 15) of gestation were placed in 1.35 mM 2dGuo, as previously described (Cowan et al., 2013). 2dGuo FTOC were then either transplanted in vivo or were used for in vitro experiments involving a further 4-d stimulation in the presence or absence of 100 µg/ml recombinant IL-4 (BioLegend) and IL-13 (PeproTech).

thymus transplantation
Freshly isolated lymphoid E18 CD45.1 + thymus lobes were transplanted under the kidney capsule of congenic CD45.2 + mice and recovered after 6 wk to analyze iNKT cells. To examine the requirement for thymic stromal expression of IL-4Rα, BALB/c WT or Il4ra −/− 2dGuotreated FTOC was transplanted into BALB/c nude mice for 10 wk. In both cases, grafting was performed as described (Cowan et al., 2013).

Quantitation of medullary epithelial-free areas
Tissue sections were randomly taken from throughout the thymus (3-4 per mouse), and the number of mTECfree areas was assessed from images taken of the whole thymus section. The size and the contribution of the areas were calculated using ZEI SS Zen Black software. To calculate areas within 100 μm of the CMJ, a line was drawn 100 μm from the CMJ.