Direct role of FLT3 in regulation of early lymphoid progenitors

Summary Given that FLT3 expression is highly restricted on lymphoid progenitors, it is possible that the established role of FLT3 in the regulation of B and T lymphopoiesis reflects its high expression and role in regulation of lymphoid‐primed multipotent progenitors (LMPPs) or common lymphoid progenitors (CLPs). We generated a Flt3 conditional knock‐out (Flt3 fl/fl) mouse model to address the direct role of FLT3 in regulation of lymphoid‐restricted progenitors, subsequent to turning on Rag1 expression, as well as potentially ontogeny‐specific roles in B and T lymphopoiesis. Our studies establish a prominent and direct role of FLT3, independently of the established role of FLT3 in regulation of LMPPs and CLPs, in regulation of fetal as well as adult early B cell progenitors, and the early thymic progenitors (ETPs) in adult mice but not in the fetus. Our findings highlight the potential benefit of targeting poor prognosis acute B‐cell progenitor leukaemia and ETP leukaemia with recurrent FLT3 mutations using clinical FLT3 inhibitors.

kinase (Matthews et al, 1991;Rosnet et al, 1991;Rosnet et al, 1993). Its ligand, FLT3 ligand (FLT3L) exists in a soluble as well as membrane-bound form (Lyman et al, 1995;Lyman & Jacobsen, 1998). Studies in mice have established that FLT3 and FLT3L play an important role in lymphopoiesis (Lyman & Jacobsen, 1998;McKenna et al, 2000;Sitnicka et al, 2002;Tsapogas et al, 2017). Although not expressed on haematopoietic stem cells (HSCs), FLT3 expression is initiated on multipotent progenitors (MPPs) and sustained on common lymphoid progenitors (CLPs), but is only expressed on the very earliest B-cell and T-cell progenitors (Wasserman et al, 1995;Adolfsson et al, 2001Adolfsson et al, , 2005Sitnicka et al, 2002;Boyer et al, 2011;Buza-Vidas et al, 2011;Luc et al, 2012). Lymphoid-primed multipotent progenitors (LMPPs) express the highest levels of FLT3 (Adolfsson et al, 2005), and FLT3 plays a key role in LMPP and CLP maintenance (Sitnicka et al, 2002. As recently highlighted (Tsapogas et al, 2017), because no Flt3 conditional knockout mouse has been generated, it remains unclear to what degree the reductions observed in B lymphocyte and thymocyte progenitors in mice with germ line deletion of FLT3 or FLT3L (Mackarehtschian et al, 1995;Sitnicka et al, 2003Sitnicka et al, , 2007, are secondary to loss of LMPPs and/or CLPs prior to becoming programmed for lymphoid-restricted development, or also reflect a distinct role of FLT3 also in already lymphoid-restricted progenitors. In fact, the expression of FLT3 in the B-and T-lymphocyte lineages, is restricted to the earliest pre-proB and early thymic progenitors (ETPs), respectively (Wasserman et al, 1995;Mansson et al, 2010;Luc et al, 2012), progenitors suggested largely to represent not fully lymphoid-restricted progenitors (Rumfelt et al, 2006;Luc et al, 2012). Establishing to what degree FLT3 plays a direct role in regulating already lymphoid-programmed progenitors, is of particular relevance for the high prevalence of two types of FLT3 driver mutations, internal tandem duplication (ITD) and recurrent FLT3 pointmutations, both associated with a poor clinical outcome in acute leukaemia (Stirewalt & Radich, 2003;Tsapogas et al, 2017), including distinct ETP and B-cell progenitor leukaemia (Armstrong et al, 2004;Neumann et al, 2013).
Furthermore, cytokine receptors and their ligands are thought to play distinct roles at different stages of development, and this has been also specifically suggested for FLT3 and FLT3L (Vosshenrich et al, 2003;Boiers et al, 2013;Beaudin et al, 2016). To more specifically address progenitor stage-and ontogeny-specific roles of FLT3 in regulation of normal lymphopoiesis, we generated a Flt3 floxed/floxed conditional knock-out (Flt3 fl/fl ) mouse line, allowing us to specifically target FLT3 deletion in a temporal and spatial manner.

Animals
The Flt3 floxed/floxed conditional knock-out (Flt3 fl/fl ) mouse line was generated using a DNA targeting construct in which the genomic fragment of the mouse Flt3 gene has exon 15 flanked by LoxP sites (flox) and with an Frt-neomycin-Frt cassette inserted into intron 15 of the mouse Flt3 gene. The IB10/C embryonic stem (ES) cell line (E14 subclone 129/Ola) was electroporated with the targeting construct and targeted clones selected using neomycin. Correctly-targeted ES clones were introduced into C57BL6 blastocysts by injection into the blastocyst cavity. Injected blastocysts were then transplanted to the uterus of pseudo-pregnant foster mothers. Offspring positive for the floxed Flt3 allele were then crossed with Flp-deleter mice to remove the neomycin cassette. Screening of Flt3 fl/fl mice was carried out using 2 primers flanking the 5 0 loxP site Primer 1: AGATGCCAGGACAT-CAGGAACCTG and Primer 2: ATCAGCCACACCAGACA-CAGAGATC. Flt3 fl/fl mice were then backcrossed for more than 5 generations with C57/Bl6 mice and subsequently crossed with different Cre-recombinase mouse strains (all on a C57/Bl6 genetic background).
Vav1 cre/+ , Mx1 cre/+ , Rag1 cre/+ mice have been previously described (Kuhn et al, 1995;McCormack et al, 2003;Stadtfeld & Graf, 2005). For each cross, non-Cre expressing Flt3 fl/fl females were bred with Flt3 fl/fl males heterozygous for the Cre of interest to yield Cre + Flt3 fl/fl as well as Cre À Flt3 fl/fl control littermates. For timed pregnancies, mice were mated late afternoon and females were checked the following morning for the presence of a vaginal plug designated as embryonic day 0Á5 (E0Á5).
All mice were maintained under specific pathogen-free conditions at Lund University Animal Facility. The Ethical Committee at Lund University approved all performed experiments.

Dissections and cell preparations
The fetal liver (FL) and fetal thymus were dissected and mechanically disrupted with a syringe. Bone marrow (BM) cells were extracted from femora and tibia using a mortar. Peritoneal cavity lavage was performed using 10 ml of phosphate-buffered saline (PBS) (Thermo Fisher Scientific Inc, Logan, UT, USA) containing 5% of Fetal Bovine Serum (FBS) (Hyclone, Logan, UT, USA). Single-cell suspensions were prepared in PBS containing 5% of FBS and filtered through a 70-lm cell strainer (BD Biosciences, San Jose, CA, USA). Cells were counted with the Sysmex (KX-21N) Haematology analyser (Sysmex Corporation Europe GmbH, Norderstedt, Germany).
Flow cytometry and fluorescence-activated cell sorting (FACS) the analysis. Samples were analysed on an LSRII (BD Biosciences) and analysis was performed using FlowJo software (version 9.3;TreeStar,Ashland,OR,USA). For all the flow cytometry profiles shown, singlet viable cells were first gated as lineage negative and further gating is indicated with arrows.

Statistics
Prism software (GraphPad Software Inc., La Jolla, CA, USA) was used for all statistical analysis. Statistical significances were determined using an unpaired Mann-Whitney test. The significance level was set at P < 0Á05.
Thus, FLT3 plays an important role in sustaining multiple stages of lympho-myeloid progenitors in adult haematopoiesis.

Requirement for FLT3 after initiation of lymphoid lineage programme
Previously reported reductions in the earliest B-and T-lymphoid progenitors in conventional Flt3 knockout mice could potentially be secondary to reductions in high FLT3-expressing LMPPs and/or CLPs rather than reflecting a specific and direct role of FLT3 in downstream B-and T-cell committed progenitors. To more specifically investigate the potential direct role of FLT3 down-stream of adult LMPPs and in already lymphoid-programmed progenitors, we crossed Flt3 fl/fl and Rag1 cre/+ mice, to exclusively target loss of FLT3 function to cells already expressing high levels of Rag1 (McCormack et al, 2003). Importantly, and in agreement with only a fraction of adult LMPPs expressing Rag1 and at very low levels (Adolfsson et al, 2005;Mansson et al, 2007Mansson et al, , 2010Luc et al, 2012), we observed no change in FLT3 cell surface expression on LMPPs, whereas a reduced fraction of Lin À SCA-1 low KIT low IL7R + CLPs expressed FLT3 (Fig 2A). Despite this, not only long-term (LT)-HSCs, short-term (ST)-HSCs, CD48 + CD150 + MPPs and CD48 + CD150 À MPPs, but also total CLP numbers were unaffected in the BM of adult Rag1 cre/+ Flt3 fl/fl mice (Fig 2B; Figure S3A), unlike in adult Vav1 cre/+ Flt3 fl/fl mice ( Figure S1B-C). The proportion of CD48 À CD150 À ST-HSCs, CD48 + CD150 + MPPs and CD48 + CD150 À MPPs, expressing FLT3 were not altered (Fig 2C-D). Notably, although the total numbers of CLPs were unaffected, the FLT3 + fraction of CLPs in Rag1 cre/+ Flt3 fl/fl mice was significantly reduced with a corresponding increase in FLT3 À CLPs (Fig 2C-D), compatible with a fraction of CLPs being sustained at normal levels although deleted for FLT3 expression. Notably, even if LMPP and CLP numbers in adult Rag1 cre/+ Flt3 fl/fl mice were unaffected, ProB, PreB and IgM + B cells in the BM (Fig 2E) as well as ETPs in the thymus were significantly reduced (Fig 2F; Figure S3B), to a similar degree as in Vav1 cre/+ Flt3 fl/fl mice ( Figure S1D-E), establishing a strict requirement for FLT3, independently of LMPPs and CLPs, and after initiation of lymphoid-restricted gene expression in adult haematopoiesis.
Previous studies have suggested that the cytokine requirement might be distinct for fetal and adult lymphoid progenitors (Carvalho et al, 2001;Vosshenrich et al, 2003;Hesslein et al, 2006;Beaudin et al, 2016;Zriwil et al, 2016). Moreover, LMPPs in the FL express higher levels of lymphoid genes, including Rag1, when compared to their adult counterparts (Boiers et al, 2013). We therefore next investigated the impact of Flt3 deletion in the E14.5 FL of Rag1 cre/+ Flt3 fl/fl embryos. In agreement with their higher Rag1 expression (Boiers et al, 2013), we observed a slight reduction in LSK cells expressing FLT3 and a more distinct and significant reduction on Lin À SCA-1 low KIT low IL-7R + CLPs (Fig 3A). Despite this, not only CD48 À CD150 + LT-HSCs, CD48 À CD150 À ST-HSCs and CD48 + CD150 + MPPs, but also CD48 + CD150 À MPPs and CLPs were unaffected in the E14.5 Rag1 cre/+ Flt3 fl/fl FL (Fig 3B). Whereas the FLT3 + and FLT3 À fractions of the different LSK HSC and MPP fractions were unaffected in the Rag1 cre/+ Flt3 fl/fl FL, FLT3 + CLPs were reduced and FLT3 À CLPs correspondingly increased (Fig 3C-D), similar to what was observed in the adult BM. Notably, the earliest ProB cell progenitors emerging in the FL at this stage, were reduced by almost 90%, demonstrating  Figure S4A-B). In contrast, the earliest thymic progenitors in the E14.5 thymus were unaffected in Rag1 cre/+ Flt3 fl/fl embryos (Fig 3F, Figure S4C). Even when E14.5 CD45.2 Rag1 cre/+ Flt3 fl/fl FL cells were competitively transplanted into lethally-irradiated adult wild-type (WT) CD45.1 recipients, we observed no impact of Flt3 deficiency on the earliest LMPP and CLP progenitor compartments (Fig 3G), while early B-and T-cell progenitors were distinctly reduced (Fig 3H-I, Figure S4D-F), again supporting a distinct role of FLT3 in early lymphoid progenitors subsequent to the initiation of lymphoid lineage programming.

Role of FLT3 in maintenance of distinct B cell subsets
As we observed a strong impact of lymphoid-restricted deletion of FLT3 on B-progenitor cell maintenance in the embryo as well as in adult haematopoiesis, we next investigated to what degree the generation and maintenance of the preferentially fetal-derived mature B1a and Marginal Zone B (MZB) cells (Hardy & Hayakawa, 1991;Kantor et al, 1992;Yoshimoto et al, 2011) as well as conventional B2 cells which are also produced during adult haematopoiesis (Hao & Rajewsky, 2001) are dependent on intact FLT3 function. At steady state, follicular B2 cells (CD19 + CD93 À CD5 À CD43 À CD23 + CD1d À ) were unaffected in the spleen and peritoneal cavity (CD19 + CD5 À CD43 À CD23 + CD11b À ) of adult Rag1 cre/+ Flt3 fl/fl mice, as were B1a (CD19 + CD93 À CD5 + CD43 + CD23 À CD1d À ) and MZB (CD19 + CD93 À CD23-CD1d + ) cells in the spleen and B1a (CD19 + CD5 + CD43 + CD23 À ) cells in the peritoneal cavity (Fig 4A-D, Figure S3C). Fetal-derived B1a cells and MZB cells are longlived and possess self-renewal potential (Hao & Rajewsky, 2001), which could result in a compensatory expansion to correct any reductions due to loss of FLT3 function. To further assess the role of FLT3 in maintenance of B1a and MZB cells, we therefore analysed the spleen and peritoneal cavity of mice transplanted with unfractionated E14.5 FL cells from CD45.2 Rag1 cre/+ Flt3 fl/fl embryos, together with competitor E14.5 FL WT CD45.1 cells. At 8 weeks following transplantation, we observed an impairment in reconstitution of B1a, MZB as well as B2 cells in the spleen and peritoneal cavity of recipients of Rag1 cre/+ Flt3 fl/fl FL cells (Fig 4E-F), establishing an important role of FLT3 for the homeostasis of each of these distinct B cell populations from already lymphoidrestricted progenitors.
To address whether FLT3 plays a direct role in the regulation of adult lymphoid-restricted progenitor maintenance, rather than in their generation from LMPPs, we specifically deleted Flt3 through Rag1 cre/+ -induced recombination (McCormack et al, 2003). As expected, due to their low levels of Rag1 expression, FLT3 expression was unaffected on LMPPs as were LMPP numbers in adult BM. Notably, whereas FLT3 expression was deleted in approximately 50% of CLPs, CLP number was also unaffected in the BM of adult Rag1 cre/+ Flt3 fl/fl mice. This allowed for the specific establishment of a distinct role of FLT3 in direct regulation of early B-and T-cell progenitors independently of the role of FLT3 in regulation of earlier LMPPs and CLPs, because the earliest BM B-cell progenitors and thymic T-cell progenitors were distinctly reduced in adult Rag1 cre/+ Flt3 fl/fl mice, in the absence of an impact on LMPPs and CLPs. This is of considerable significance, given that FLT3 expression in the B and T cell lineages is restricted to the very earliest B-and T-cell progenitors, pre-proB cells (Wasserman et al, 1995) and ETPs (Luc et al, 2012), respectively. Importantly, these progenitors are likely to also be key cellular targets for recurrent FLT3 driver mutations in patients with acute B-cell progenitor and ETP leukaemia (Carow et al, 1996;Armstrong et al, 2004;Neumann et al, 2013). Our findings therefore highlight the potential benefit of targeting these leukaemias with clinical FLT3 inhibitors (Annesley & Brown, 2014).
Also in the liver of Rag1 cre/+ Flt3 fl/fl embryos, we observed a dramatic reduction in early B-cell progenitors in the absence of a significant reduction of fetal LMPPs and CLPs, demonstrating a distinct and prominent requirement for FLT3 in fetal B-lymphoid progenitors, again independently of a role in earlier progenitors. Interestingly, distinct from adult haematopoiesis, E14.5 fetal thymic progenitor homeostasis was not affected in Rag1 cre/+ Flt3 fl/fl mice, despite FLT3 being expressed on fetal ETPs (Luis et al, 2016). However, at 4 weeks after competitive transplantation of Rag1 cre/+ Flt3 fl/fl FL cells into adult recipients, while CD48 + CD150 À MPPs and CLPs remained unaffected, an impaired generation of not only B cell progenitors but also early thymocyte progenitors could be readily detected from Rag1 cre/+ Flt3 fl/fl FL cells. The distinct thymocyte phenotype observed upon transplantation of E14.5 Rag1 cre/+ Flt3 fl/fl FL cells into adult recipients but not in the Rag1 cre/+ Flt3 fl/fl E14.5 fetal thymus, combined with a similar early thymocyte defect observed in unperturbed adult Rag1 cre/+ Flt3 fl/fl mice, suggest a distinct requirement for FLT3 signalling in adult but not fetal thymopoiesis, potentially explained by distinct differences in the fetal and adult thymic environment. In contrast, fetal and adult Blymphopoiesis are both dependent on intact FLT3 expression and function.
Despite its very restricted expression, our studies demonstrate that FLT3 plays an important and distinct role in the direct regulation of early B cell-restricted progenitors in FL as well as in adult BM. In further agreement with its crucial role in B-lymphopoiesis, our studies of Rag1 cre/+ Flt3 fl/fl mice establish a critical role for FLT3 in the maintenance of fetal-derived B1a and MZB cells (Hardy & Hayakawa, 1991;Kantor et al, 1992;Yoshimoto et al, 2011) as well as conventional B2 cells (Hao & Rajewsky, 2001). Previous studies have shown that the important role of FLT3 in fetal B1 and B2 lymphopoiesis is even more evident in the absence of interleukin 7 (IL7). Specifically, in the absence of the IL7 ligand and FLT3 ligand, fetal and adult B lymphopoiesis is almost entirely lost, suggesting that thymic stromal lymphoprotein (TSLP), also acting through the IL7 receptor, is unable to rescue B cell development in the absence of FLT3 ligand and IL7 (Sitnicka et al, 2003;Jensen et al, 2008).
In conclusion, through lymphoid-restricted and ontogenyspecific deletion of FLT3 expression and function, we establish that the important role of FLT3 in fetal and adult B-and T-lymphopoiesis is, in fact, not primarily explained by its role in regulation of LMPPs and CLPs, but rather by a direct and more prominent role in the regulation of the very earliest Rag1 expressing B-and T-cell progenitors, which are likely to be primary cellular targets for recurrent FLT3 mutations in clinically distinct B-and T-cell progenitor cell leukaemia. Representative fluorescence-activated cell sorting (FACS)profiles of Follicular B2 cells (CD19 + CD93 À CD5 À CD43 À CD23 + CD1d À ), Marginal Zone B cells (MZB: CD19 + CD93 À CD23-CD1d + ) and B1a cells (CD19 + CD93 À CD5 + CD43 + CD23 À CD1d À ) in spleen in 12-week-old Rag1 +/+ Flt3 fl/fl and Rag1 cre/+ Flt3 fl/fl mice (numbers in gates represent percentages of total spleen cells). (B) Mean percentages (AESD of total spleen cells) of follicular B2, MZB and B1a cells in 12-week-old Rag1 +/+ Flt3 fl/fl and Rag1 cre/+ Flt3 fl/fl mice (n = 6 mice per genotype in 2 experiments). (C) Representative FACS profiles of B2 cells (CD19 + CD5 À CD43 À CD23 + CD11b À ) and B1a cells (CD19 + CD5 + CD43 + CD23 À ) in the peritoneal cavity (PC) of 12-week-old Rag1 +/+ Flt3 fl/fl and Rag1 cre/+ Flt3 fl/fl mice (numbers in gates represent percentages of total PC cells). (D) Mean percentages (AESD of total PC cells) B2 cells and B1a cells of 12-week-old Rag1 +/+ Flt3 fl/fl and Rag1 cre/+ Flt3 fl/fl mice (n = 6 mice per genotype in 2 experiments). (E-F) Mean percentages (AESD) contribution of CD45.2 cells to (E) follicular B2 cells, MZB cells and B1a cells in the spleen and to (F) B2 cells and B1a cells in the PC of CD45.1 wild type (WT) mice transplanted with 2 9 10 6 unfractionated E14.5 FL cells from Rag1 +/+ Flt3 fl/fl (CD45.2) or Rag1 cre/+ Flt3 fl/fl (CD45.2) mice (n = 2-5 per genotype) together with 2 9 10 6 cells unfractionated E14.5 FL WT CD45.1 competitor cells, analysed 8 weeks post-transplantation. Author contributions AZ, ES and SEWJ designed and conceptualized the overall research and analysed the data. CN designed and supervised the generation of the Flt3 conditional knockout targeting construct and targeting of ES cells. AZ performed the experiments. TAK performed B1 cell analysis experiments. LW provided expertise in the animal work. CB contributed to the design, analysis of experiments, data analysis and writing of the manuscript. JY contributed with expert advice and input on B cell development. AZ, ES and SEWJ wrote the manuscript. All authors read and approved the submitted manuscript. and MC_UU_12009/5 (SEWJ) and an International recruitment award from the Swedish Research Council (SEWJ). CB was supported by a postdoctoral fellow grant from the Swedish Childhood Cancer Foundation (PDS13/005) and the Swedish Research Council, Marie Sklodowska Curie Actions, Cofund, Project INCA (#2015-00135, #600398). ES had an Associate Professor position supported by the Swedish Childhood Cancer Foundation (TFJ08).

Declaration of interest
The authors declare no competing financial interests.

Supporting Information
Additional supporting information may be found online in the Supporting Information section at the end of the article. Figure S1. Provides data validating the model. Figure S2. Provides data that extends the findings in Fig 1. Figure S3. Provides data that extends the findings in Figs 2, 3 and 4. Figure S4. Provides data that extends the findings in Fig 3. Table S1 . Provides the list of antibodies used in the study.