SUMMARY
Multiple waves of hematopoietic progenitors with distinct lineage potentials are differentially regulated in time and space. The site of origin and the cellular outputs of the first wave of thymic seeding progenitors remain unclear. Here we show that these progenitors are T/ILC restricted and unique in generating lymphoid tissue inducer cells, in addition to embryonic-derived invariant Vγ5+ γδT cells that are both essential for the development of thymic medullary cells and thus shape the thymic architecture. We further show that all lymphoid progenitors are exclusively of hematopoietic stem cell origin, despite the temporal overlap between thymopoiesis initiation and the transient expression of lymphoid transcripts in yolk sac precursors. This early transient expression does not alter their strict erythro-myeloid differentiation potential. Our work highlights the relevance of the developmental timing on the emergence of different lymphoid subsets required for the establishment of a functionally diverse immune system.
INTRODUCTION
Three lymphoid subsets develop exclusively during embryonic life, they comprise T lymphocytes expressing Vγ5 that reside in the skin, also named dendritic epidermal-like T cells (DETC), Vγ6 that reside in the lungs and in the genital urinary tract and lymphoid tissue inducers (LTis), a subset of group 3 innate lymphoid cells (ILC3). Both Vγ5+ T cells and LTis are required for the development of medullary thymic epithelial cells (mTECs), through TNFRSF signaling, a mechanism which is only conserved in prenatal stages (Alves et al., 2014; Boyden et al., 2008; Roberts et al., 2012; Rossi et al., 2007). LTis are also required for secondary lymphoid tissue organization (Cupedo and Mebius, 2003; Cupedo et al., 2002) whereas Vγ6 contributes to remodeling of the tissues they colonize (Papotto et al., 2017).
During embryogenesis, the thymus is colonized by two waves of hematopoietic progenitors named thymic seeding progenitors (TSP), with distinct potentials (Cumano et al., 2019; Ramond et al., 2013). The first wave of TSP follow a unique developmental program as they share phenotype, transcriptional signature and restricted T cell differentiation potential with HSAlo α4β7- fetal liver (FL) common lymphoid progenitors (CLPs) (Berthault et al., 2017; Ramond et al., 2013). These CLPs are no longer found after E15.5, a stage after which the thymus is colonized by a second wave of TSPs that share phenotype, transcriptional signature and differentiation potential with lympho-myeloid primed progenitors (LMPPs) (Berthault et al., 2017; Bhandoola et al., 2007; Luc et al., 2012; Ramond et al., 2013). Because the first TSP that colonize the thymus were the only to generate Vγ5+ T cells (Ikuta et al., 1990; Ramond et al., 2013), it was important to determine whether they also generate thymic LTis, that together shape the thymic architecture (Roberts et al., 2012). Accordingly, ILC3 were the majority of embryonic thymic ILCs, gradually lost after birth, while a small population of ILC2 is maintained throughout life (Jones et al., 2018). It remained also to be determined whether the first wave of TSP are yolk sac (YS) or hematopoietic stem cell (HSC) derived. Addressing this question highlights a critical challenge in understanding T cell function.
Several successive but overlapping waves of emerging hematopoietic progenitors, with different lineage potential, are differentially regulated in time and space (Cumano and Godin, 2007; Perdiguero and Geissmann, 2016). Hematopoietic cells first appear in the extra-embryonic yolk sac (YS) blood islands at around embryonic day 7-7.5 (E7-7.5), where primitive erythropoiesis occurs prior to the emergence of multi-lineage erythro-myeloid progenitors (EMPs) (Bertrand, 2005; Palis et al., 1999). The emergence of pre-HSCs starts at E9.5 in the aorta-gonad-mesonephros (AGM) and major arteries (Bertrand et al., 2010; Boisset et al., 2010; de Bruijn et al., 2000; Kissa and Herbomel, 2010), colonize the FL around E10.5 and later (around E16) migrate to the bone marrow (BM) (Cumano and Godin, 2007; Dzierzak and Bigas, 2018).
HSC are distinguished from all other progenitors by their ability to long-term repopulate all hematopoietic lineages (Cumano and Godin, 2007; Cumano et al., 2001). Emerging HSC cannot however be tested in assays similar to those used for adult HSC and are designated as pre-HSC (Cumano et al., 2001; Medvinsky and Dzierzak, 1996). Most hematopoietic cells are constantly differentiating from HSCs although a few lineages are HSC independent and exclusively produced during embryonic development. For example, YS cells that contribute to erythropoiesis and megakaryopoiesis are also the source of tissue resident macrophages that persist throughout life (Bertrand, 2005; Gomez Perdiguero et al., 2015a; Hoeffel et al., 2015; McGrath et al., 2015). This heterogeneity and layered organization of the hematopoietic system lead to the possibility that, similar to tissue macrophages, the first innate-like lymphocytes could also be YS derived. In line with this view, cells identified as lympho-myeloid restricted progenitors (LMPs) were observed in E9.5 YS before HSC activity is detected. They expressed lymphoid associated genes (Il7r, Rag2, Rag1) and were proposed as the origin of the first wave of TSP (Luis et al., 2016). Independent reports have also converged to support the notion that Vγ5+ and another γδ T cell subset, Vγ6+ IL-17 producing might originate from YS progenitors, independent of HSCs (Gentek et al., 2018a; Spidale et al., 2018). In contrast to the above studies, Vγ5+ and B1a B cells were shown to be preferentially derived from a particular HSC-like subset, transiently found in the FL but not in the adult BM, marked by an history of Flk2 expression (Beaudin et al., 2016).
In this report, we undertook a characterization of the functional properties and fate of the progeny of the first TSP as well as their developmental origin. Unlike the current view that during embryogenesis ILCs are derived exclusively from the common α4β7+ ILCs progenitors (Chea et al., 2016; De Obaldia and Bhandoola, 2015; Possot et al., 2011), we found that the first wave of TSP is composed of bipotent T/ILC progenitors that contribute to the thymic LTi cells, highlighting their unique and unprecedentedly reported role in generating ILC. Moreover, the first TSP-derived conventional T cells that lack terminal deoxynucleotidyl transferase (TdT) expression and therefore exhibit a restricted T cell repertoire, are rapidly overridden by subsequent waves of T cells. Therefore, the major role of the first wave of TSP might be restricted to provide cells that drive thymic organogenesis and homeostasis (Abramson and Anderson, 2017; Cumano et al., 2019; White et al., 2018). We further demonstrated for the first time in vivo that the embryonic TSPs are not YS-derived but rather originate from intra-embryonic HSCs. Interestingly, YS-derived progenitors although showing a history of lymphoid-associated gene expression, such as Il7r, Rag2, Rag1, did not differentiate into any lymphoid subset in vivo or in vitro. This transient expression, also found in other FL myeloid progenitors, is restricted to embryonic hematopoiesis and uncoupled from differentiation potential. Altogether our data highlight the impact of embryonic developmental timing on lymphocyte production, gene expression and the heterogeneity of the immune system.
RESULTS
Initial TSP show LTi biased lineage potential in vitro
A detailed analysis of the transcriptional profiles of TSP from the first (E13) and the second (E18) waves identified the overexpression, in the former, of ILC related transcripts from which stood out LTi-associated lineage transcripts (Rorc, Cd4, Cxcr5, Il1r1 and Ltb); Supplementary Fig 1a) (Ishizuka et al., 2016; Ramond et al., 2013). This comparison also showed that genes involved γδ T cells development (Bhlhe40, Cited4, B4galnt2, Tgm2 and Txk) were differentially expressed between E13 and E18 TSPs (Supplementary Fig 1a). Of note, Sell and Ly6a were upregulated in E18 TSP, in line with their resemblance to LMPPs (Berthault et al., 2017; Ramond et al., 2013). In addition, the E14 DN3 thymocytes were devoid of TdT activity (Supplementary Fig 1b) giving rise to T cells with a restricted T cell repertoire raising the possibility that they do not contribute to the adaptive T cell compartment, but rather to the innate-like T cell and ILC compartments.
To determine their developmental potential, we performed a two-step culture (Berthault et al., 2017; Luc et al., 2012) (Fig 1a) that allowed the generation of all major lymphoid lineages (T, B, ILC, NK) and myeloid cells from single progenitors (Fig 1b). Consistent with our previous observations showing that TSP before E15.5 are restricted lymphoid progenitors biased for T cell generation (Ramond et al., 2013), we found that E13 TSP only generated T cells and ILC whereas E18 TSPs retained B and myeloid potential. Most TSP gave rise to T and ILCs, although a few (6-15%) only generated T cells and none gave rise to ILC only. About 59% of the clones derived from E18 TSP generated all subsets of ILC and NK cells, whereas only 24% of E13 TSP showed that capacity, suggesting that the latter are more differentiated and restricted in their potential than the former (Fig 1b, c).
The most striking difference between the two subsets was that 22% of E13 TSP and none of the E18 TSP generated only ILC3 in addition to T cells (Fig 1c). These results indicated that mid gestation (E12.5-15.5) TSP appear LTi biased.
The initial TSP show LTi-associated gene expression
To understand the molecular basis of the ILC lineage potential differences between E13 and E18 TSP we analyzed by multiplex qRT-PCR, the expression of 41 lymphoid-associated genes in single cells. Unsupervised hierarchical clustering identified 4 distinct clusters (Fig 2). Cluster I comprised a majority of E18 TSP and showed no expression of ILC associated transcripts, consistent with their capacity to generate all ILC populations. Clusters II and IV were essentially composed of E13 TSP and were characterized by the expression of common T/ILC genes Tox, Tcf7 and Gata3. Cluster IV also expressed the LTi cell associated transcripts Cxcr6, Cd4, Rorc, Cxcr5, Il1r1 and Ltb, consistent with their ILC3 restricted differentiation potential, whereas Cluster II expressed the innate-like T cells/ILC-associated transcripts Zbtb16, Gzmb, Nfil3, Itga2b and Gzma (De Obaldia and Bhandoola, 2015). Cluster III contained two thirds of E13 TSP and one third of E18 TSP and was characterized by the expression of low levels of Tox, Gata3, Tcf7, Sox4, Runx1 and Runx3 indicative of common T/ILC priming and consistent with a broad ILC differentiation potential (Fig1c).
Most TSP showed reduced signs of Notch signaling activation with few cells expressing Hes1 consistent with their in vitro capacity to generate both T cells and ILC. This is in line with the observation that FL lymphoid progenitors expressed ILC associated transcripts before settling the thymus (Berthault et al., 2017; Possot et al., 2011).
Embryonic TSP have multi-ILC lineage potential in vivo
We next assessed the ILC lineage potential of E13 and E18 TSPs in vivo. CD45.2+ E13 or E18 ETPs were intravenously transferred into 1-day old Rag2-/-γc-/- CD45.1+ mice. Recipient mice were analyzed at 5 weeks post-transfer (Fig. 3a). While mice engrafted with E13 TSP only generated ILC, E18 TSP also generated B and myeloid cells, (Fig 3b; Supplementary Fig 2b,c) consistent with the more restricted differentiation potential of E13 TSP. As expected, few T cells were generated in these chimeras because of the atrophic thymus in Rag2-/-γc-/- mice and because TSP down-regulate Ccr7 and Ccr9 as they enter the thymus and consequently lose their ability to home back to the thymus (Zlotoff et al., 2010).
Analysis of intestinal lamina propria, where all ILC lineages coexist, showed that TSP from both waves generated NK cells (EOMES+), ILC1 (EOMES-), ILC3 (RORɣt+) and ILC2 (GATA3+). In the lung we found ILC2, in the spleen NK cells and in the liver NK cells and ILC1 (Fig 3c,d; Supplementary Fig 2a). These results indicated that E13 and E18 TSP have the capacity to generate all ILC subsets in vivo and that the ILC generated recapitulated the expected tissue distribution.
Neonatal thymectomy has no effect on tissue colonization by embryonic derived Vγ5+ and Vγ6+ γδ T cells
Vγ5+ and Vγ6+ γδ T cells in the skin and in the lymph nodes (LN), respectively, are of embryonic origin (Haas et al., 2012; Ikuta et al., 1990). To assess the dependency of these embryonic tissue resident γδ T subsets on thymic output that might inform on their origin in the first or second wave of TSP, we performed neonatal thymectomy. Newborn (up to 36 hours after birth) mice were thymectomized or sham-operated as a control (Fig 4a). At 6 weeks of age complete thymectomy was ascertained by the absence of double positive CD4/CD8 cells in the tissue around the thymus location (Fig 4b). Thymectomized mice (n=6) were severely lymphopenic as shown by the numbers of CD4 and CD8 αβ T cells in the inguinal (i) LN (Fig 4d) and in the spleen; of note, CD44- αβ T cells in the spleen were severely reduced (Fig 4e), a sign of activation of the T cell compartment consequent to the lymphopenia and consistent with the absence of a functional thymus in these mice. By contrast, the numbers of Vγ5+ γδ T cells in the skin and those of Vγ4- IL17+ γδ T cells (considered as Vγ6+ γδ T) in the iLN were similar in sham and in thymectomized animals (Fig 4c, d). By contrast, iLN Vγ4+ IL17+ γδ T cells were severely reduced (Fig 4d). These experiments suggest that the two embryonic γδ T cell subsets are likely derived from the first wave of TSP. All other αβ or γδ T depend on post-natal thymic output and thus are likely largely issued from the second wave of TSP.
E9.5 YFP expressing Il7rαCreRosa26YFP YS-progenitors are devoid of lymphoid potential
The experiments above indicated that the first cells that colonize the thymus are particularly efficient in generating ILC and embryonic γδ T cells. Because YS derived progenitors express lymphoid associated genes such as Rag, Flt3 and Il7r (Böiers et al., 2013; Luis et al., 2016) we considered the possibility that the first TSP originated in an HSC independent pathway. This notion was further reinforced by a recent report suggesting a YS origin of Vγ5 + gd T (Gentek et al., 2018a).
We used a lineage tracer mouse line to identify the onset of Il7r-expressing cells, Il7rαCreRosa26YFP (Schlenner et al., 2010). We performed time course analyses of YFP expressing cells in the YS, head, AGM region and FL (Fig 5a). No YFP expression was observed in E8.5 embryos, in any of the analyzed tissues in either the Kit+ or Kit- fraction (data not shown). Consistent with previous studies, YFP+ progenitors were first detected in Lin- Kit+CD41+ YS cells and were undetected in the AGM or in the placenta (Fig 5d) at E9.5 in Il7raCreRosa26YFP embryos (Fig 5b,c). 5-8 % of the c-kit+ cells in the YS were YFP+ (Fig 5b,c), also expressed CD31 and CD16/32 and lacked surface expression of CD135, CD127 or CD115 (Fig 5e), the phenotype of YS EMPs (Bertrand, 2005; McGrath et al., 2015).
To further investigate the potential of YS YFP+ cells to develop into T/GM and B/GM cells, single cells were cultured into monolayers of OP9-DL4 or OP9 stroma (Fig 5f). YFP+ YS cells showed myeloid potential and lacked detectable T or B cell potential (Fig 5g, h). Methylcellulose-based colony assays showed that both Lin-Kit+YFP+ and Lin-Kit+YFP- YS cells exhibited comparable GM and MkE potential (Fig 5i). These results indicated that expression of YFP in the E9.5 YS Il7raCreRosa26YFP embryos does not equate with lymphoid potential, and that both YFP+ and YFP- E9.5 YS Kit+ cells showed a lineage potential similar to erythro-myeloid progenitors (EMPs) (Bertrand, 2005; Gomez Perdiguero et al., 2015a; McGrath et al., 2015).
To determine whether E9.5 YS Kit+ progenitors could give rise to the first differentiating γδ T cells, the Vγ5+ T cells, we performed FTOC of thymic lobes colonized with either E9.5 YS Kit+ YFP+ progenitors, E10.5 YS-Kit+ progenitors or E12 FL LSKs as a control (Fig 5j). Consistent with their inability to generate T cells in OP9-DL4 cultures, E9.5 YS-derived cells did not generate Vγ5+ T cells or initiate thymopoiesis. By contrast, E10.5 YS-Kit+ progenitors and E12 FL LSK readily generated T cells including Vγ5+ T cells. E10.5 YS-Kit+ progenitor derived T cells originate from AGM-derived pre-HSC, known to enter circulation between E9.5 and E11.5 (Delassus and Cumano, 1996; McGrath et al., 2003) and were used to probe the sensitivity of the T cell assay. These experiments showed that E9.5 YS progenitors are devoid of lymphoid potential under culture conditions sufficiently sensitive to detect few circulating T cell progenitors.
Expression of Il7r and Flk2 in YFP+ E9.5 IL7rCreRosaYFP YS progenitors is transient
To better characterize YFP+ YS progenitors, which share phenotype and differentiation potential with EMPs, we analyzed the expression of transcripts associated with multiple hematopoietic lineages in single E9.5 YS YFP+ and YFP- Lin- CD117+ CD41+ cells. Unsupervised hierarchical clustering segregate YFP+ and YFP- cells into four clusters (Supplementary Fig 3a). Clusters I and II, comprises mostly YFP+ cells and are characterized by the expression of Pu-1, Irf8 Csf1r a pattern of expression corresponding to EMPs (Mass et al., 2016) (Supplementary Fig 3b). Cluster II and a fraction of cluster I showed also expression of, Kdr, Tek, Runx1, Mpl and Foxo1, suggesting that they were EMP recently specified from hemogenic endothelium (Baron et al., 2018; Kasaai et al., 2017; Mass et al., 2016). Furthermore, cells in cluster II also expressed the lymphoid-associated genes Il7ra and Flk2. Clusters III and IV comprised both YFP- and YFP+ cells expressing high levels of Gata1 indicating their erythroid lineage engagement and did not express any of the lymphoid genes. Cells in cluster IV also expressed Klf1 indicating that they are more advanced into the erythroid lineage differentiation. Ebf1, Pax5 or Rag1 were not detected in any YS cells Quantitative PCR performed on bulk populations showed that YFP+ cells expressed significantly higher levels of Rag1, Rag2 and Il7r than YFP- cells albeit 10-fold lower than CLP (Supplementary Fig 4). Altogether, these experiments indicated that the transient lymphoid associated gene expression in YS cells is restricted to the more immature cells, i.e. the emerging YS-derived EMP and is dissociated from their differentiation potential.
FL myeloid progenitors are labelled in Il7raCreRosa26YFP mouse embryos
To study the possibility that YS derived progenitors contribute to the lymphoid lineage during embryogenesis in vivo, we mapped YFP+ cells in Il7raCreRosa26YFP mice. We found that, unlike in adult BM where Il7rCre labels exclusively lymphoid progenitors (Schlenner et al., 2010), YFP expression in the FL was also found in multipotent and myeloid progenitors (Supplementary Fig 5a,b). FL and fetal blood (FB), short-term HSCs (ST-HSCs), multipotent progenitors (MPPs), granulocytes-myeloid progenitors (GMPs), common myeloid progenitors (CMPs) and megakaryocytes-erythroid progenitors (MEPs) were labeled at frequencies ranging from 10%-40% (Supplementary Fig 5a,b; see Supplementary Figure 6 for gating strategy). The frequency of YFP+ ST-HSC and myeloid progenitors progressively decreased after birth and was undetectable in adult mice (Schlenner et al., 2010) (Supplementary Fig 5b). FL YFP+ GMPs and CMPs lacked any detectable B cell potential in vitro, retaining GM potential (Supplementary Fig 5c). These observations indicated that at early stages of development Il7raCre also labels cells devoid of lymphoid potential, and therefore this model does not faithfully trace the origin of embryonic lymphoid cells.
To establish whether YFP+ cells contribute to tissue-specific hematopoietic lineages known to be of embryonic origin (Gentek et al., 2018b; Gomez Perdiguero et al., 2015b) we analyzed the brain microglia, the FL Kupffer cells and the skin mast and Langerhans cells of E18 Il7raCreRosa26YFP. We found expression of YFP in about 40% of all these populations (Supplementary Fig 7a, b).
Thymopoiesis-initiating cells develop exclusively from HSC-derived progenitors
To define the relative contribution of YS cells and HSCs to the first wave of TSPs, we used a fate mapping mouse model expressing the tamoxifen inducible Cre (MeriCreMer) under the control of the Csf1r promoter (Qian et al., 2011) (Fig 6a). We crossed Csf1rMeriCreMer mice with the Cre reporter mouse strain Rosa26YFP or Rosa26Tomato (based on the availability of fluorochrome-labelled antibodies for the analysis) and induced recombination in embryos by a single injection of Hydroxytamoxifen (4-OH-TAM) into E8.5, E9.5 or E10.5 pregnant females. Csf1r is first expressed in E8.5 YS progenitors (Gomez Perdiguero et al., 2015a), later it is expressed in E9.5 YS Il7raCreRosa26YFP (Supplementary Fig 3) (Böiers et al., 2013) and at E10.5 Csf1r is also expressed in pre-HSC (Baron et al., 2018) (Supplementary Fig 8d). Consequently, single 4-OH-TAM injection at E8.5 or E9.5 should mark Csf1r-expressing myeloid cells and EMP of YS origin that also comprise Il7r traced cells. Injection of 4-OH-TAM at E10.5 should mark HSC derived progenitors and Csf1r-expressing myeloid cells.
In embryos pulsed at E8.5, YFP+ cells were detected in the E9.5 YS. These cells expressed Csf1r, Il7r and were devoid of lymphoid potential (Supplementary Fig 8a,b,c). In embryos pulsed at either E8.5 or E9.5 and analyzed at E12.5, YFP+ HSC, HSC-derived lymphoid progenitors and TSP were either not detectable or represented less than 1% in FL and thymus, respectively (Fig 6b,c). In contrast, 30-50% of microglia cells were labeled, in line with their YS origin. Embryos pulsed at E10.5 and analyzed at E12.5 showed significant labeling in populations of FL HSC, multipotent, lymphoid and myeloid progenitors (Fig 6c). Interestingly, the frequency of Tomato+ cells in E12.5 TSP (25%) and in FL LTi (22%) (See supplementary Fig 8e for gating strategy) paralleled that of HSCs (31%).
These fate-mapping experiments directly demonstrated that no detectable TSP and lymphoid progenitor originate from YS cells under physiological conditions; rather their origin coincides in time with that of pre-HSC.
Innate-like Vγ5+ T cells or B1-a cells that differentiate during a limited time window of embryonic development have been proposed to derive from HSC-independent precursors (Gentek et al., 2018a; Ghosn et al., 2016). Therefore, we analyzed Vγ5+ T cells and B-1a in the skin and peritoneal cavity respectively of adult Csf1rMeriCreMer Rosa26YFP or Csf1rMeriCreMer Rosa26Tomato mice pulsed at either E8.5, E9.5 or E10.5 (See supplementary Fig 8f for gating strategy).
Adult mice pulsed at E8.5 showed no YFP labelled lymphocytes and those pulsed at E9.5 showed less than 1% of Vγ5+ T or B-1a labelled cells, while 30-40% of microglia cells were YFP labelled (Fig 6c). By contrast, embryos pulsed at E10.5 showed frequencies of Vγ5+ T or B-1a labelled cells similar to that found in HSCs (Fig 6c). These results are consistent with a strict HSC origin of all lymphoid cells including the first wave of TSP and innate-like lymphocytes of embryonic origin.
DISSCUSION
Here we show that the first wave of embryonic TSP developed exclusively from HSC and show a unique capacity to generate LTi and Vγ5+ γδ T cells that, together, shape the thymic architecture.
The interaction between the first developing Vγ5+ T cells and CD4+ CD3- LTi with thymic medullary epithelial cells, results in the expression of Aire by the latter, which is essential to establish thymic T cell tolerance (Abramson and Anderson, 2017; Roberts et al., 2012; Rossi et al., 2007). We have previously shown that the first TSP derive from a unique and transient set of hematopoietic progenitors, T-biased in their differentiation potential, that colonize the thymus between E12.5 and E15.5 (Berthault et al., 2017; Ramond et al., 2013). These precursors were the only capable of generating Vγ5+ γδ T cells (Ramond et al., 2013). We show here, by single cell lineage potential assays, that these TSP comprise a population restricted to the T and ILC3 lineages. This observation was further supported by transcriptional analyses showing a cluster of E13.5 TSP that, although not committed, are primed and biased towards the ILC3 lineage with high expression of Rorc, Cd4, Cxcr5, Il1r1 and Ltb, a signature of LTi (Ishizuka et al., 2016; Kernfeld et al., 2018). Together with the lack of detectable ILC restricted progenitors within E13.5 thymocytes and the temporally linked appearance of thymic LTi and Aire+ mTECs (Rossi et al., 2007), these results indicate that the first TSP contribute to both Vγ5 T cells and thymic LTi and are therefore the main actors in shaping thymic architecture.
The first TSP were also devoid of TdT activity giving rise to T cells without N sequences and thus with a restricted T cell repertoire. Low TdT expression together with the poor proliferative capacity and the observation that after P4 virtually all αβ expressing conventional T cells have N sequences in their V(D)J junctions (Bogue et al., 1992), suggested that the contribution of the first TSP to the conventional T cell compartment is limited (Cumano et al., 2019).
In vivo TSP were capable of restoring the tissue resident ILC compartments of newborn Rag2γc-/- mice. This finding suggests that TSP in circulation can home to the peripheral tissues and differentiate in situ into all ILC populations.
Our data show that the first TSP isolated using generally accepted markers (Porritt et al., 2004; Ramond et al., 2013) are T/ILC restricted with poor B and myeloid potential arguing against the notion that the embryonic thymus is initially seeded by lympho-myeloid restricted progenitors. Using Rag1-GFP mice, it was argued that E11.5 TSP retained myeloid potential and resembled PIRA/B+ LMPPs in phenotype, lineage potential and molecular signature (Luis et al., 2016). However, it is not clear whether these cells were TSP because, at this stage, the epithelium that forms the thymus anlage is not yet folded resulting in a difficult distinction from the surrounding pharyngeal pouches. Moreover, the analysis based on the unique expression of GFP might include a heterogeneous population of circulating hematopoietic progenitors, thus compromising the conclusions from this study. Different strategies to isolate TSP might explain the discrepancies in lineage potential and molecular signature between the two studies.
As the origin of progenitors during development may be linked to the generation of distinct populations within the immune system, we revisited the controversial origin of the first wave of TSP. By analogy with the documented YS origin of tissue resident macrophages (Gomez Perdiguero et al., 2015b; Perdiguero and Geissmann, 2016), it was proposed that also early developing innate-like T cells and therefore the first TSP were HSC independent (Böiers et al., 2013; Gentek et al., 2018a; Kobayashi et al., 2014; Luis et al., 2016; Spidale et al., 2018; Yoshimoto et al., 2011, 2012).
The evidence that YS is a potential source of early lymphocytes is based on adoptive transfer experiments, on mice that lack blood circulation or by tracing YS cells expressing lymphoid-associated genes (Böiers et al., 2013; Yoshimoto et al., 2012). The complexity of hematopoietic cell generation that occurs in multiple sites in overlapping time windows hampers the precise assignment of hematopoietic progenitor origin. To unambiguously assign the contribution of each anatomical site to blood cell formation, faithful lineage tracing models that trace the progeny of each source in a non-overlapping spatio-temporal manner have to be used.
Our analysis of YFP expressing cells under the control of the Il7r regulatory sequences throughout embryonic life indicated that, unlike adult hematopoietic progenitors, multipotent and myeloid restricted embryonic progenitors express lymphoid associated genes, albeit in a transient manner, without any consequent restriction in their differentiation potential. Consistent with our analysis, a recent report using the same mouse model demonstrated that FL hematopoietic progenitors and tissue resident macrophages were significantly labeled (Gautier et al., 2012; Heng et al., 2008; Leung et al., 2019). These observations established that in hematopoiesis, gene expression does not necessarily equate with lineage potential.
The analyses of inducible lineage tracing of Cdh5 and Runx1 expressing cells (Gentek et al., 2018a) suggested that Vγ5+ T cells were generated from early emerging hematopoietic cells, precisely from YS-derived hematopoietic progenitors. Because Cdh5 and Runx1 are expressed in both endothelial and hematopoietic cells these models do not allow a clear discrimination between progenitors originated in the YS or in the dorsal aorta. Moreover, the lack of direct comparison between the labeling efficiency in HSC and in Vγ5+ T cells compromise the conclusions obtained from these mouse models.
Csf1rMerCreMer lineage tracing allowed a temporal analysis and unambiguously traced YS derived progenitors (including those expressing lymphoid-associated genes) in females induced at E8.5 and E9.5, while it traced both HSC and YS derived Csf1r expressing myeloid cells, when pulsed at E10.5.
Microglia but virtually no HSC, Vγ5+ γδ T cells or B1a B cells were labeled by a single pulse at E8.5. In contrast, induction at E10.5 labelled HSC, Vγ5+ T cells and B1a B cells with similar efficiencies, indicating a lineage relationship. Moreover, when pulsed at E9.5 a low (less than 1%) frequency of labelled HSC was accompanied by equally low frequency of labelled innate-like lymphocytes. These fate mapping experiments are the first in vivo evidence that directly demonstrate that TSPs, innate-like lymphocytes and lymphoid progenitors originate in the embryo from Csf1r+ HSCs-derived progeny, show no contribution of HSC-independent progenitors to lymphopoiesis and settle the controversy over the lymphoid ontogeny during embryogenesis. Together these data indicate that lymphopoiesis is a hallmark of HSC and HSC-derived progenitors
The existence of restricted lymphoid progenitors in the FL between E11.5 to E15.5 suggests that the generation of embryonic-derived tissue resident lymphoid cells follows a unique developmental program. These cells are produced prior to the establishment of a fully competent adaptive immunity and persist throughout life, that possibly contribute to the integrity of body barriers and act as a first line of defense against pathogens. The ability of embryonic TSP to generate ILCs is the first evidence that they can originate also from hematopoietic progenitors other than the classical ILCP, and in organs other than the FL where hematopoiesis occurs (Fig 7). This highly orchestrated sequence of events highlights the requirement of a layered immune cell production whereby the first wave of TSPs is essential for thymic organogenesis and homeostasis.
AUTHOR CONTRIBUTIONS
R.E. designed and performed most experiments, analyzed data and wrote the manuscript; S.M. analyzed chimeric mice and contributed to discussions. F.S.S. and T.P. performed Biomark analysis. O.B.-D performed neonatal thymectomy. E.G.P., L.F. and L.I. performed fate-mapping experiments. A.B. performed adoptive transfer in neonatal mice. R.G., P.V., A.B. and P.P. contributed to the discussions. A.C. directed the research, designed experiments, analyzed data and wrote the manuscript. All authors contributed to the manuscript.
COMPETING FINANCIAL INTERESTS
The authors declare no competing financial interests.
Methods
Animals
Mice were bred in dedicated facilities of the Institut Pasteur under specific pathogen-free conditions. Mattings were done overnight; on the following day, mice showing a vaginal plug were isolated and were considered to be embryonic day 0 (E0). Il7rCre (Schlenner et al., 2010), Rag2-/-; γc-/- CD45.1 (Colucci et al., 1999), ROSA26Tomato (Madisen et al., 2010) and ROSA26YFP mice were on C57BL/6 background, Csf1rMeriCreMer (Qian et al., 2011) mice were on FVB background. All animal manipulations were performed according to the ethic charter approved by French Agriculture ministry and to the European Parliament Directive 2010/63/EU.
Fate-mapping of Il7r+ hematopoietic progenitors
For fate-mapping analysis of Il7r+ progenitors, Il7rCre females were crossed to homozygous ROSA26YFP. Indicated embryonic tissues and adult tissues were analyzed by flow cytometry.
In utero pulse labelling of Csf1r+ Progenitors
For fate-mapping analysis of Csf1r+ progenitors, we crossed Csf1rMeriCreMer with either ROSA26YFP or ROSA26Tomato reporter mice. For induction of reporter recombination in the offspring, a single dose of 75 µg per g (body weight) of 4-Hydroxytamoxifen (Sigma) (4OHT) supplemented with 37.5 µg per g (body weight) progesterone (Sigma) (to counteract the mixed estrogen agonist effects of 4OHT, which can result in fetal abortions) was injected either at E8.5 or E9.5 (Gomez Perdiguero et al., 2015). To induce recombination in the embryos at E10.5, females received a single dose of 1.2 mg 4OHT and 0.6 mg progesterone.
Cell suspension
Cell suspensions from embryonic, newborn and adult tissues were obtained through mechanical dissociation and/or enzymatic digestion. In brief, all cells were suspended in Hanks’ balanced-salt solution (HBSS) supplemented with 1% FCS (GIBCO). Embryonic tissues were micro dissected under a binocular magnifying lens. Organs were rinsed with HBSS plus 1% FCS to remove blood cells contamination. Then, thymus and fetal liver were passed through a 26-gauge needle of a 1-ml syringe and were filtered. Blood from embryos was obtained for 20 to 30 min HBSS supplemented with 1% FCS without calcium and magnesium. Blood cells were resuspended in a 70% (vol/vol), topped by a 40% (vol/vol), solutions of Percoll. After 20 min of centrifugation at 3000 r.p.m, cells were collected at the 40%-70% interface. For embryonic skin samples, harvested back skin was minced with scissors in RPMI medium containing 0.1 mg/ml DNase I (Roche) and 1mg/ml Collagenase I (GIBCO). Digestions were performed for 30-45 min at 37°C under continuous agitation. For embryonic head/brain, tissues were cut into small pieces, incubated in RPMI containing 10% FCS and 0.2 mg/ml Collagenase IV (GIBCO) for 30 min and then passed through a 19G needle to obtain a homogenous cell suspension. Adult thymi and spleen were dissected, and single cell suspension were obtained through a nylon mesh. Bone marrow were collected by flushing bones with HBSS. For adult skin, ears were harvested from 8-10 week-old mice, the dorsal and ventral sheets were separated using forceps and the epidermis was removed by incubation for 45 min at 37°C in 2.4 mg/ml of Dispase II (Invitrogen) and 3% FCS. Then, the epidermis was further digested for 30 min in PBS 0.1 mg/ml DNase I (Roche) and 1mg/ml Collagenase I (GIBCO). Adult brain was processed similar to embryonic brain except for an incubation period of 1 hour and cell suspension was resuspended in a 70% (vol/vol) solution of Percoll topped by a 40% (vol/vol) solution of Percoll. After 20 min of centrifugation at 3000 r.p.m, cells were collected at the 40%-70% interface.
For reconstituted Rag2γc-/- tissue preparations, animals were exsanguinated, and the lung was minced and incubated for 30 min at 37°C with agitation in RPMI containing 2% FCS, then incubated for 1 hour with RPMI medium containing 0.2 mg/ml Collagenase IV (GIBCO) and 0.1 mg/ml DNase I (Roche). Cells were resuspended in a 70% (vol/vol) solution of Percoll topped by a 40% (vol/vol) solution of Percoll. After 20 min of centrifugation at 3000 r.p.m, cells were collected at the 40%-70% interface. Small intestines were minced into small pieces and digested in RPMI medium containing 0.1 mg/ml DNase I (Roche) and 100 mg/ml Collagenase VIII (Sigma). Cell suspension was resuspended in a 70% (vol/vol) solution of Percoll topped by a 40% (vol/vol) solution of Percoll. After 20 min of centrifugation at 3000 r.p.m, cells were collected at the 40%-70% interface. Livers were teased and cells were resuspended in RPMI with 2% FCS, centrifuged for 7 min, then further purified after 20 min of centrifugation at 3000 in 44% Percoll. Peritoneal washouts of adult mice were performed by injecting IP PBS with 2% FCS and collecting the resulting peritoneal fluid.
Flow cytometry and cell sorting
All digested samples were filtered, and single cell suspensions were blocked with anti-CD16/32 (BD Biosciences) and stained with antibodies listed in supplementary table 1 and fixable viability dye (Invitrogen). For transcription factor staining, cells were stained for surface markers for 30 min at 4°C then fixed according to manufacturer’s instructions (eBiosciences) and stained with fluorescent antibodies to intracellular proteins. Stained cells were analyzed on a LSR Fortessa flow cytometry or sorted using a FACSAria III (BD Biosciences). Data were analyzed with FlowJo software (Treestar).
Culture. All experiments were done in 96-well plates at 37°C and 5% CO2 and in complete OptiMEM medium supplemented with 10% FCS, penicillin (50 units/ml), streptomycin (50 µg/ml) and β-mercaptoethanol (50 µM) (GIBCO).
a) Frequency assay. Single cells were sorted into 96-well plates containing a monolayer of 1000 OP9 (for B cell-ILC and myeloid potential) or OP9-DL4 (for T cell potential) stromal cells in complete medium supplemented with saturating amounts of IL-7, Flt3L, IL-2 and KitL for B, T and ILC differentiation and M-CSF, GM-CSF and KitL for myeloid differentiation. For erythroid, megakaryocytic and myeloid potential, cells were directly sorted into wells of 96-well plate and were supplemented with KitL, erythropoietin for erythrocytes differentiation, thrombopoietin for megakaryocytes differentiation, M-CSF and GM-CSF for myeloid differentiation. Cultures were supplemented with fresh cytokines every 7 d. After 12 days of culture, wells showing colonies were stained for CD19 (B cells), NK1.1 (NK cells), CD4 (T cells), CD8 (T cells), CD3 (T cells), CD11b (Myeloid cells), Gr-1(Myeloid), CD41 (megakaryocytes) and Ter119 (erythrocytes) or analyze ILC-associated transcription factors after cell fixation, GATA-3 (ILC2), RORgt (ILC3) and Eomes-(ILC1). Frequency scores were assigned based on the frequency of positive wells relative to the total number of plated cells.
b) Single cell T cell-ILC-B cell-myeloid-potential assay. Single cells were directly sorted onto a monolayer of OP9 stromal cells in Terasaki plates in medium supplemented with IL-7, IL-2, KitL and Flt3L. After 36h, growing cells were split into OP9 or OP9-DL4 cultures under the same conditions mentioned above, and clones were analyzed after 12 days.
Reconstituted FTOC
Irradiated thymic lobes (30Gy) from CD45.1+ mouse embryos at E14 or E15 were colonized for 48h, on a hanging drop, in Terasaki plates by 200 YS-derived progenitors from CD45.1+ embryos at E9.5, or E10.5 or 200 LSKs from CD45.1+ embryos at E12.5, in culture medium. After colonization, thymic lobes were cultured for 12d on a filter (Millipore) floating on 3ml of complete medium in a 37°C incubator with 5% CO2. Then, thymi were dissociated and analyzed by flow cytometry.
In vivo adoptive transfer
800-1200 fetal E13 HSCs, E13 and E18 ETPs from CD45.2+ embryos were sorted using FACSAria III, then injected intravenously into irradiated (1.5G) 1 day old Rag2-/-; γc-/- CD45.1+ mice. Recipient mice were analyzed 5 weeks post-transfer.
Neonatal Thymectomy
Thymectomy was performed on anesthetized mice within 36 hours of birth. The sternum was split and the bilobed thymus removed by mechanical suction and the skin closed with surgical adhesive. After 6 weeks, the skin, the thymus, the inguinal lymph nodes and the spleen were analyzed. Controls were animals of the same age were sham operated without thymectomy.
Quantitative RT-PCR
mRNA of sorted cells was extracted using the RNeasy Micro Kit (Qiagen), and cDNA was obtained using the PrimeScript RT Kit (Takara). qPCR was performed using TaqMan primers and TaqMan Universal Master Mix (Applied Biosystems) for the following genes: Il7r, Rag1, Rag2, Csf1r, Csf2r, Csf3r and Myb. Gene expression was normalized to that of Hprt and relative expression was calculated using the 2-ΔCt method.
Microarray
Microarray data in the GEO database with the accession code GSE50910 published by Ramond et al (Ramond et al., 2013).
Multiplex single-cell qPCR
Single cells were sorted in 96-well PCR plates in 9 µl of RT/pre-amp mix from the CellsDirect One-Step qRT-PCR Kit (Life Technologies) and were kept at −80°C at least overnight. For each subset analyzed, a control-well with 20 cells were sorted. Pre-amplified cDNA (20 cycles) was obtained according to manufacturer’s note and was diluted 1:5 in TE buffer, pH 8 (Ambion) for qPCR. Multiplex qPCR was performed using the microfluidics Biomark system on Biomark HD for 40 cycles (Fluidigm). TaqMan probes that were used for RT/pre-amp and qPCR are listed in supplementary table 3. Only single cells for which the three housekeeping genes could be detected before 20 cycles were included in the analysis. Analysis was done in R package Phenograph as in (Perchet et al., 2018)
Statistical analysis
Statistical data show either mean or mean ± s.e.m., and the two tailed unpaired Student’s t-test was used. Statistic tests were performed using Prism software.
ACKNOWLEDGMENTS
We thank S. Novault, S. Megharba and S. Schmutz from the flow cytometry core facility of the Institut Pasteur for technical support; the staff of the animal facility of the Institut Pasteur for mouse care. We thank H.-R. Rodewald for providing Il7raCreRosa26YFP mice and S.E.W. Jacobsen for fruitful discussions. This work was financed by the Institut Pasteur, INSERM, Pasteur-Weizmann Foundation and ANR (grant Twothyme) through grants to A.C.; by REVIVE (Investissement d’Avenir; ANR-10-LABX-73) to A.C. and R.E.; by recurrent funding from the Institut Pasteur, the CNRS, Revive (Investissement d’Avenir; ANR-10-LABX-73) and by an ERC investigator award (2016-StG-715320) from the European Research Council to E.G.P.