University of Birmingham Formation of the Intrathymic Dendritic Cell Pool Requires CCL21-Mediated Recruitment of CCR7+ Progenitors to the Thymus

During ab T cell development in the thymus, migration of newly selected CD4 + and CD8 + thymocytes into medullary areas enables tolerance mechanisms to purge the newly selected ab TCR repertoire of autoreactive speciﬁcities. Thymic dendritic cells (DC) play key roles in this process and consist of three distinct subsets that differ in their developmental origins. Thus, plasmacytoid DC and Sirp a + conventional DC type 2 are extrathymically derived and enter into the thymus via their respective expression of the chemokine receptors CCR9 and CCR2. In contrast, although Sirp a 2 conventional DC type 1 (cDC1) are known to arise intrathymically from immature progenitors, the precise nature of such thymus-colonizing progenitors and the mechanisms controlling their thymus entry are unclear. In this article, we report a selective reduction in thymic cDC1 in mice lacking the chemokine receptor CCR7. In addition, we show that the thymus contains a CD11c + MHC class II 2 Sirp a 2 Flt3 + cDC progenitor population that expresses CCR7, and that migration of these cells to the thymus is impaired in Ccr7 2 / 2 mice. Moreover, thymic cDC1 defects in Ccr7 2 / 2 mice are mirrored in plt/plt mice, with further analysis of mice individually lacking the CCR7 ligands CCL21Ser ( Ccl21a 2 / 2 ) or CCL19 ( Ccl19 2 / 2 ) demonstrating an essential role for CCR7-CCL21Ser during intrathymic cDC1 development. Collectively, our data support a mechanism in which CCR7-CCL21Ser interactions guide the migration of cDC progenitors to the thymus for correct formation of the intrathymic cDC1 pool. The Journal of the controlling the via its ligand CCL21Ser.

T he ability of ab T cells to recognize foreign Ags presented by self-MHC complexes takes place during T cell development in the thymus (1)(2)(3). Positive selection of immature CD4 + CD8 + thymocytes triggers the expression of multiple chemokine receptors, including CCR4 and CCR7 (4,5), enabling newly produced CD4 + and CD8 + single-positive thymocytes to migrate toward the thymus medulla (6), where T cell tolerance mechanisms take place. At this site, CD4 + and CD8 + thymocytes are screened for high-affinity ab TCR recognition of self-antigens, including those controlled by Aire (2,7,8). Although high-affinity ab TCR signaling in CD8 + thymocytes results in negative selection by apoptosis, CD4 + thymocytes can undergo two fates: either negative selection or diversion into the Foxp3 + T regulatory (Treg) lineage (9)(10)(11). Consequently, intrathymic elimination of selfreactive thymocytes biases conventional thymic T cell production toward self-tolerant cells, whereas Foxp3 + Treg development limits the autoimmune potential of developing T cells that escape thymic selection events. In the thymus medulla, specialized stromal microenvironments contain both medullary thymic epithelial cells (mTEC) and dendritic cells (DC) that express MHC class I and MHC class II (MHC II) and act as effective APCs (7,12,13). Importantly, ab TCR screening for both negative selection and Foxp3 + Treg development can take place via direct recognition of self-antigens on mTEC themselves or following the transfer Ags to thymic DC (14)(15)(16)(17). Thus, interplay between mTEC and DC in the thymus is important to maximize opportunities for self-antigen recognition during T cell tolerance induction.
Thymic DC are heterogeneous, consisting of both extrathymically and intrathymically derived populations (18,19). Thus, both Sirpa + conventional DC type 2 (cDC2) (20)] and plasmacytoid DC (pDC) are generated extrathymically, with their entry into the thymus providing a source of self-antigens from peripheral tissues. In contrast, Sirpa 2 conventional DC type 1 (cDC1) arise from immature progenitors that colonize the thymus and complete their maturation intrathymically (21,22) prior to acting as APC for mTEC-derived self-antigens (15,23). Thus, intrathymic DC availability depends upon multiple DC subtypes. Consequently, the mechanisms that control their relative contributions to the intrathymic DC pool are important in understanding thymic tolerance. For example, thymic Sirpa + cDC2 express the chemokine receptor CCR2 and are reduced in Ccr2 2/2 mice that display defects in negative selection (24). In addition, pDC are CCR9 + , and Ccr9 2/2 mice show defects in the recruitment of pDC to the thymus and are impaired in thymocyte deletion (25). Moreover, ligands for both CCR2 (CCL2) and CCR9 (CCL25) are expressed by thymic stromal cells (26)(27)(28), highlighting the importance of thymic microenviroments in the control of thymic DC. Although these studies are important, as they explain how pDC and cDC2 are able to contribute to the intrathymic DC pool, the mechanisms that control intrathymic Sirpa 2 cDC1 (20) are less clear. Indeed, although several studies have studied DC potential within thymic cells (21,22,29,30), direct examination of the mechanisms regulating thymic cDC1 has been hindered by limitations in the identification of cDC-committed progenitors in the thymus. In contrast, stages of cDC development in peripheral lymphoid tissues are well defined (20,31), and cDCcommitted progenitors (pre-cDC) have been identified at multiple sites, including spleen and bone marrow (32)(33)(34). For example, in cell transfer experiments analyzing splenic DC development, pre-cDC with a Lin 2 CD11c + MHC II 2 Flt3 + Sirpa low phenotype were shown to selectively give rise to cDC progeny but not pDC or monocytes (32). Importantly, however, although such pre-cDC have been identified in peripheral tissues, their presence in the thymus has not been examined. Consequently, mechanisms regulating the entry of DC progenitors into the thymus, and the possible requirement for particular chemokine receptors in this process, have not been addressed.
In this study, we have examined development of the intrathymic DC pool in the adult mouse thymus. We find that the thymus contains a population of Lin 2 CD11c + MHC II 2 Flt3 + Sirpa low pre-cDC that expresses the chemokine receptor CCR7. In adult Ccr7 2/2 mice, we show that a selective reduction in cDC1 correlates with a reduction in thymic pre-cDC, with short-term in vivo homing assays indicating a reduced ability of Ccr7 2/2 pre-cDC to enter the thymus. Finally, by analyzing mice lacking expression of individual CCR7 ligands, we demonstrate a selective reduction in thymic pre-cDC and DC1 in CCL21Ser-deficient (Ccl21a 2/2 ) but not CCL19-deficient (Ccl19 2/2 ) mice. Collectively, our study demonstrates a mechanism in which CCR7 regulates thymic cDC1 development by controlling the intrathymic availability of pre-cDC via its ligand CCL21Ser.
Tracking DC migration in vivo using fluorescent microbeads Short-term tracking of DC migration in vivo was performed exactly as described (25). In brief, 200 ml of yellow/green (YG) fluorescent (505/515) carboxylate-modified microspheres (FluoSpheres, 0.2 mm diameter; Invitrogen) were i.v. injected into adult WT or Ccr7 2/2 mice. Forty-eight hours postinjection, thymus and spleen tissues were isolated, and beadlabeled DC subsets and pre-cDC were analyzed by flow cytometry.

Proliferation analysis using BrdU
BrdU incorporation was used to detect proliferation of cDC. A total of 1.5 mg BrdU was injected i.p. into mice, which were sacrificed 18 h later. Thymic cell suspensions were prepared by enzymatic digestion, and cDC1 and cDC2 populations were identified as described above. To reveal BrdU incorporation, cells were permeabilized and stained using the APC BrdU Flow Kit according to the specification (BD Pharmingen).

Statistical analysis
All analyses used GraphPad Prism 6.0. Statistical analysis was performed using unpaired Student t tests. Only p values ,0.05 were identified as significant. Nonsignificant differences were not highlighted. In all figures, error bars represent SEM.

CCR7 controls intrathymic availability of Sirpa 2 cDC1 and their progenitors
Although chemokine receptors are known to play important roles in the recruitment of peripheral cDC2 and pDC to the thymus (24,25,39), mechanisms that establish intrathymic cDC1 from immature thymuscolonizing progenitors are less clear. Given that CCR7 and its ligands play an important role in the migration of DC in peripheral lymphoid tissues (40)(41)(42), we first examined the intrathymic DC pool in Ccr7 2/2 mice. Thymus and spleen cell suspensions from adult WT and Ccr7 2/2 mice were prepared, and Lineage 2 (Lin 2 ) CD11c + PDCA1 2 cDC were identified by flow cytometry (Fig. 1A). Both the proportion and absolute number of intrathymic cDC were significantly reduced in Ccr7 2/2 mice ( Fig. 1A-C). Further subdivision of total thymic cDC using Sirpa to identify Sirpa 2 cDC1 and Sirpa + cDC2 revealed that there was a significant reduction in cDC1 numbers (Fig. 1B, 1C). Importantly, splenic cDC1 proportions and numbers were comparable in WT and Ccr7 2/2 mice (Fig. 1D, 1E), arguing against a systemic loss of these cells in the absence of CCR7. Interestingly, cDC2 numbers were comparable in the thymus of WT and Ccr7 2/2 mice (Fig. 1C), indicating that the mechanisms controlling cDC2 entry to the thymus are not limited by CCR7 deficiency. In contrast, the selective cDC1 reduction in the thymus of Ccr7 2/2 mice suggests that CCR7 is required for the thymic entry of these cells or their progenitors. In support of this, analysis of intrathymic DC populations following in vivo BrdU administration demonstrated comparable proportions of BrdU + cDC1 in both WT and Ccr7 2/2 thymus ( Fig. 2A, 2B), indicating that reduced thymic cDC numbers in Ccr7 2/2 mice are not due to reduced cell proliferation.
Although pre-cDC have been defined in peripheral lymphoid tissues (32)(33)(34), the precise nature of corresponding DC progenitors in thymus is still not fully clear. For example, the presence of Lin 2 CD11c + MHC II 2 Flt3 + Sirpa low pre-cDC (32) in the thymus has not been studied, and the relationship between these cells and other thymic DC progenitors described in additional studies is not fully clear (21,22,43). Interestingly, we found that Lin 2 CD11c + MHC II 2 Flt3 + Sirpa low pre-cDC were readily detectable in the thymus of adult WT mice (Fig. 3A), albeit at a lower frequency compared with both spleen and bone marrow (Fig. 3B). To see whether the reduction in intrathymic cDC1 in Ccr7 2/2 mice correlated with alterations in the frequency of pre-cDC, we first used flow cytometric analysis and anti-CCR7 Abs to examine CCR7 expression on thymic DC subsets. In agreement with earlier reports (21, 44, 45), we found that thymic cDC1 and cDC2 both expressed CCR7 (data not shown). Interestingly, pre-cDC in the thymus were also CCR7 + , with higher CCR7 levels detectable on thymic pre-cDC, as compared with bone marrow pre-cDC (Fig. 3C). In addition, although pre-cDC numbers were comparable in the bone marrow of WT and Ccr7 2/2 mice (Fig. 3D), we saw a significant reduction in pre-cDC in the thymus of Ccr7 2/2 mice (Fig. 3D). Thus, our findings indicate that the selective loss of cDC1 in the thymus of Ccr7 2/2 mice is accompanied by a reduction in numbers of intrathymic CCR7 + pre-cDC, suggesting a role for CCR7 in the recruitment of these cells to the thymus. Furthermore, that pre-cDC are present at normal frequency in bone marrow also indicates that their reduction in the thymus is not likely due to limited availability caused by alterations in pre-cDC development at extrathymic sites.
Thymic recruitment of pre-cDC is impaired in Ccr7 2/2 mice CCR7 plays an important role in the migration of newly selected CD4 + and CD8 + thymocytes into the thymus medulla (6), and the absence of CCR7 or its ligands results in disrupted medulla organization and small medullary areas (5,46). To examine whether the defects in cDC1 and pre-cDC in Ccr7 2/2 mice are secondary to these alterations in medulla size, we generated bone marrow chimeras using mixtures of CD45.1 + WT and CD45.2 + Ccr7 2/2 progenitors, in which WT haemopoietic cells restore thymic medulla architecture (6). As controls, we established similar chimeras using mixtures of congenically marked CD45.1 + WT and CD45.2 + WT bone marrow, and all cells were transferred into CD45.1 + CD45.2 + lethally irradiated hosts to allow identification of transferred WT and Ccr7 2/2 progeny (Fig. 4A). Mice were harvested after 8 wk, and anti-CD45.1/anti-CD45.2 Abs were used to examine chimerism within thymic cDC and pre-cDC populations. As expected, the contribution of each donor to total thymus cellularity was comparable in both WT:WT and WT:Ccr7 2/2 chimeras (Fig. 4B). Moreover, WT and Ccr7 2/2 bone marrow showed comparable contributions to intrathymic cDC2 in WT: Ccr7 2/2 chimeras (Fig. 4C). In contrast, we saw a significant decrease in the proportion of cDC1 generated from Ccr7 2/2 bone marrow in WT:Ccr7 2/2 chimeras (Fig. 4D). Moreover, this reduction in intrathymic cDC1 generated from Ccr7 2/2 marrow was accompanied by a significant reduction in the proportion of Ccr7 2/2 -derived pre-cDC (Fig. 4E). Thus, reductions in cDC1 and pre-cDC in unmanipulated Ccr7 2/2 mice still occur in the presence of WT counterparts, indicating these effects are not secondary to medulla disorganization in Ccr7 2/2 mice.
To directly examine the recruitment of pre-cDC to the thymus in the steady-state, we adopted a short-term homing assay used previously to examine pDC entry to the thymus, in which migratory DC are labeled by uptake of fluorescent microbeads (25). Thus, WT mice were i.v. injected with YG-labeled microbeads, and splenic and thymic DC populations were analyzed for YG labeling 2 d postinjection. As expected following i.v. transfer into WT mice, YG + cells were clearly detectable within all cDC1, cDC2, and pre-cDC populations in the spleen (Fig. 5A, 5C). Interestingly, we saw differential labeling of DC populations in the thymus. Thus, ∼10% of intrathymic cDC2 were YG + (Fig. 5B, 5C), consistent with the extrathymic origin of these cells. In contrast, very few (1-2%) of intrathymic cDC1 were labeled YG + following i.v. microbead injection. This low frequency of bead uptake by intrathymic cDC1 is in line with their intrathymic generation and is also indicative that this labeling approach does not readily label thymic DC in situ,  perhaps because of the blood-thymus barrier (47). Importantly, analysis of intrathymic pre-cDC in the same mice showed that ∼10% of these cells were YG + , indicating their migration to the thymus from peripheral sites (Fig. 5A, 5C). Next, when we compared YG-labeled DC populations in tissues from WT and Ccr7 2/2 mice after i.v. microbead transfer, we saw a significant reduction in the numbers of both YG + pre-cDC and cDC1 in the thymus of Ccr7 2/2 mice (Fig. 5D). This was not due to differential cell labeling between strains, as no differences in the numbers of YG + pre-cDC and cDC1 were seen in the spleens of WT and Ccr7 2/2 mice (Fig. 5E). Moreover, and consistent with unaltered cDC2 numbers in Ccr7 2/2 mice (Fig. 1), numbers of YG + cDC2 in the thymus of WT and Ccr7 2/2 mice were comparable (Fig. 5D). Thus, by tracking the steady-state migration of DC subsets using shortterm in vivo homing assays, our findings indicate that Lin 2 CD11c + MHC II 2 Flt3 + Sirpa low pre-cDC enter the thymus from the periphery and that this process is reduced in the absence of CCR7.
CCL21, but not CCL19, controls intrathymic DC pool formation CCL19, CCL21Ser, and CCL21Leu represent the three known functional chemokine ligands for CCR7 (48). As the genes encoding both CCL19 (Ccl19) and CCL21Ser (Ccl21a) are expressed in multiple thymic stromal cell types (49-51), we next investigated whether the requirement for CCR7 in intrathymic cDC1 development mapped to specific chemokine ligand requirements. Initially, we examined intrathymic DC in plt/plt mice, in which expression of both Ccl19 and Ccl21a is absent (35). Thus, freshly isolated thymus tissue from adult WT and plt/plt mice was enzymatically digested, and intrathymic Sirpa 2 cDC1 and Sirpa + cDC2 DC subsets were identified by flow cytometry.
Consistent with the requirement for CCR7 and the phenotype of Ccr7 2/2 mice, plt/plt mice showed a reduction in the absolute numbers of total thymic cDC and a selective reduction in the proportion and numbers of cDC1 (Fig. 6A-C). In addition, we also saw a significant reduction in the absolute numbers of pre-cDC in plt/plt mice (Fig. 6C). Thus, for intrathymic DC populations, plt/ plt mice essentially mirror the effects seen in Ccr7 2/2 mice. Next, to examine the requirement for individual CCR7 ligands in thymic DC development, we examined Ccl19 2/2 and Ccl21a 2/2 mice that individually lack expression of the CCR7 ligands CCL19 or CCL21Ser. Interestingly, no alterations in the numbers and proportions of cDC1, cDC2, and pre-cDC were seen in the thymus of Ccl19 2/2 mice (Fig. 6A, 6B, 6D). In contrast, analysis of Ccl21a 2/2 mice showed alterations in thymic DC populations. In particular, we saw a reduction in the number of total cDC (Fig. 6E) that was caused by a specific reduction in both the proportion and number of cDC1 (Fig. 6A, 6B, 6E). Moreover, pre-DC were also reduced in the thymuses of Ccl21a 2/2 mice (Fig. 6E). Thus, analysis of mice that lack CCR7 ligands either individually or in combination indicates that although CCL19 is dispensable, CCL21Ser plays an essential role in controlling the intrathymic development of cDC1.

Discussion
In the thymus medulla, interactions between mTEC, DC, and newly selected thymocytes are essential for both negative selection and Foxp3 + Treg generation, which represent key aspects of T cell tolerance. Although thymic DC are known to be heterogeneous, the mechanisms that control formation of the intrathymic DC pool from its constituent components of peripherally derived pDC and cDC2 and intrathymically produced cDC1 are unclear. Given that chemokine receptors play important roles in the thymic recruitment of pDC and cDC2 (24,25), we investigated their potential role in the development of cDC1 in the thymus. In particular, given the role of CCR7 in both thymocyte migration and DC migration in peripheral lymphoid tissues, we examined the role of this chemokine receptor and its ligands during development of the intrathymic DC pool.
In this article, we show that the thymus of Ccr7 2/2 mice has a selective defect in the frequency of cDC1. Mixed bone marrow chimeras show this defect maps to CCR7 expression by haemopoietic cells and is not an indirect consequence of the medullary disorganization seen in these mice. Furthermore, we show that a pre-cDC subset, previously described only in peripheral lymphoid tissues (32), is present in the thymus and expresses CCR7. Moreover, such pre-cDC are reduced in the thymus of Ccr7 2/2 mice, with in vivo migration assays indicating this deficiency is caused by their reduced capacity to enter the thymus. Thus, the contribution of cDC1 to the intrathymic DC pool occurs via a mechanism involving CCR7-mediated recruitment of pre-cDC. This requirement for CCR7 by cDC1 draws parallels with the respective requirements of pDC for CCR9 and cDC2 for CCR2  The Journal of Immunology (24,25) and extends our understanding of the importance of chemokine receptors in thymic DC development. In addition, our finding that thymic cDC1 are altered in CCl21a 2/2 but not Ccl19 2/2 mice highlights the importance of individual chemokines for thymic DC, with the CCR7 ligand CCL21Ser playing an essential role in intrathymic cDC1 development. Recently, another study reported that increased DC apoptosis in Ccr7 2/2 mice resulted in alterations in their intrathymic DC (44). However, although this study did not examine DC progenitors and mechanisms of their thymus entry, it is important to note that the cDC1 deficiency we describe in this article is accompanied by a reduction in intrathymic CCR7 + pre-cDC. Also, during intrathymic DC development, DC progenitors may downregulate CCR7 as they progress to an immature MHC II low DC stage, which is followed by phases of steady-state maturation in the thymus that involve MHC II upregulation and the reexpression of CCR7 (45). Taken together, these observations suggest that CCR7 may play multiple roles at different stages during thymic DC development, including colonization by migrant DC progenitors, and subsequent intrathymic survival and/or maturation of their cDC1 progeny. This scenario is perhaps similar to the multiple roles played by CCR7 during conventional ab T cell development in the adult thymus that include lymphoid progenitor colonization and cortex-to-medulla migration of positively selected thymocytes (52)(53)(54)(55). In addition, that the reduction in pre-cDC in Ccr7 2/2 mice does not result in increased compensatory proliferation in either these cells or their cDC1 progeny (data not shown) may also indicate a limited availability of intrathymic growth factors for DC and/or their progenitors.
Our findings are also significant in relation to the intrathymic developmental potential of pre-cDC defined by a Lin 2 CD11c + MHC II 2 Flt3 + Sirpa low phenotype. For example, when such pre-cDC were isolated from bone marrow and transferred i.v., both cDC1 and cDC2 progeny were detectable in the spleen of recipient mice (32). Thus, our finding that the reduction in thymic pre-cDC in Ccr7 2/2 mice results in a selective deficiency in cDC1, but not cDC2, appears at odds with their capacity to act as common progenitors for cDC. One possible explanation is that as cDC2 can enter the thymus from the periphery as already mature cells, these cells then occupy a finite number of appropriate niches (43). This may then limit the intrathymic generation of cDC2 from colonizing pre-cDC, which may result in their intrathymic skewing toward cDC1 development. Alternatively, pre-cDC that enter the thymus may represent a particular subset of these cells that may already be biased toward cDC1 development. Further comparative analysis of DC progenitors in thymus and peripheral lymphoid tissues may help in discriminating these possibilities. It is also interesting to note that although cDC1 and cDC2 both express CCR7 (21,44), the cDC defect in Ccr7 2/2 mice maps to cDC1 and not cDC2. Thus, redundancy in the chemokine receptors expressed by cDC2 (e.g., CCR2) may still promote their efficient migration to the thymus. In contrast, CCR7 appears to represent a dominant chemokine receptor for intrathymic cDC1 development, although the presence of at least some pre-cDC/cDC1 in the thymus of Ccr7 2/2 mice may also suggest compensatory roles for other chemokine receptors, albeit less effectively than CCR7.
In addition, by analyzing the chemokine ligand requirements of thymic DC, we show that CCL21Ser is both essential and sufficient for CCR7-mediated control of thymic cDC1 and their progenitors. Interestingly, that thymic DC require CCL21Ser but not CCL19 may be similar to the requirements of DC in lymph nodes, where DC homeostasis and function were reported to be unaffected in Ccl19 2/2 mice (56). It is also worthy to note that in the thymus, CCL21Ser expression has recently been shown to map specifically to mTEC (38), which also control the intrathymic positioning of cDC1 via their expression of XCL1 (57). Taken together, such findings emphasize the importance of mTEC in the regulation of thymic DC and highlight roles for multiple chemokines in both the recruitment (CCL21Ser) and intrathymic positioning (XCL1) processes that take place during thymic cDC1 development. Interestingly, however, although mTEC expression of CCL21 is controlled by LTbR signaling (50), absence of LTbR expression by TEC does not perturb thymic DC numbers (58). Thus, additional receptors expressed by the thymic epithelium may also trigger CCL21 expression to regulate intrathymic cDC. Finally, the paucity of thymic cDC1 in Ccr7 2/2 and Ccl21a 2/2 mice described in this article may also be important in explaining the importance of CCR7 and its ligands in central tolerance. Indeed, CCL21Ser, but not CCL19, has recently been shown to be important for T cell tolerance in the thymus, where it controls medulla entry of positively selected thymocytes (38). Taken together, these findings indicate that CCR7-CCL21Ser interactions may be important for central tolerance in two separate ways: regulation of , compared with controls (black bars). All analysis was obtained from a minimum of nine mice per strain across at least three independent experiments. Error bars represent the SEM using an unpaired Student two-tailed t test. *p , 0.05, **p , 0.01, ***p , 0.001, ****p , 0.0001. thymocyte cortex-to-medulla migration and the regulation of thymic cDC1 availability. In conclusion, our study shows that CCR7 determines cDC1 development in the thymus via a mechanism involving its ligand CCL21Ser and the recruitment of CCR7-expressing pre-cDC. These findings highlight the importance of multiple chemokine receptors in controlling the makeup of the intrathymic DC pool and demonstrate further the key influence of CCR7 on thymus function.