Fetal Lymphocyte Progenitors Efficiently Generate Innate-Like B Cells but Make a More Limited Contribution to the Production of Conventional Lymphocytes

Abstract

Various progenitor populations with lymphoid potential emerge in multiple waves of fetal development and can do so independent from hematopoietic stem cells (HSCs). This review compares and contrasts data from our analysis of a mutant mouse in which fetal but not adult lymphopoiesis is intact with results from lineage tracing studies. This allowed us to address how efficiently and how long fetal lymphoid progenitors generated lymphoid progeny after birth. We then present, based on the data in these reports, a working model proposing that fetal lymphocyte progenitors efficiently generate innate-like B-1 and Marginal Zone B cells but their contribution to the production of B and T cells of the adaptive immune system is more limited. Instead, lymphocyte development from adult HSCs is necessary to fill secondary lymphoid tissues with conventional lymphocytes and fetal progenitors lack the potential to do so. We propose that, in addition to generating innate-like effectors, the key role of fetal lymphoid progenitors is to produce a sufficient number of B and T cell progeny, particularly in the intestines, so that the newborn and neonate have functional immunity until lymphopoiesis from hematopoietic stem cells is fully established.

TEASER ABSTRACT

Progenitors with lymphoid potential emerge in multiple waves of fetal development independent from hematopoietic stem cells. We compare and contrast data obtained from analysis of a mutant mouse in which fetal but not adult lymphopoiesis is intact with results from lineage tracing studies and present a model proposing that fetal progenitors efficiently generate innate-like B cells. However, their contribution to the production of conventional B and T cells is limited and is instead dependent on adult HSCs. We propose that the key role of fetal lymphoid progenitors is to produce sufficient lymphocytes so that the newborn and neonate have functional immunity until lymphopoiesis from hematopoietic stem cells is fully established.

INTRODUCTION

The traditional view of blood cell development is that hematopoietic stem cells (HSCs) arise in utero, seed the bone marrow, and continuously generate all B and T cells present in adult tissues after birth. However, it is increasingly apparent that this linear view of lymphocyte development is not correct. Instead, it is now recognized that various progenitors with lymphoid potential emerge in multiple, waves of fetal development and can do so independent from stem cells [1-4]. The initial part of this review will summarize work from several laboratories that established this paradigm.

We then synthesize data from our laboratory [5] with the findings from recent lineage tracing studies [3, 4] and present a model proposing that most innate-like B lymphocytes in adult lymphoid tissues are the progeny of fetal lymphoid progenitors while the contribution of the latter cells to other B and T cell compartments is more limited.

LYMPHOCYTE PROGENITORS EMERGE IN MULTIPLE WAVES IN UTERO

The Herzenberg’s first proposed that several types of progenitors evolved sequentially and functioned at specific times in the fetus and adult to generate the lymphocytes that comprise the adult immune system [6]. This layered immune system hypothesis was formulated in part to understand data indicating that innate-like B-1 B cells were most efficiently generated from fetal progenitors [7, 8]. B-1 B cells are found in the highest concentration in serous cavities [7, 9] and classical B-1 B cells and B-1 derived plasma cells are the source of spontaneously secreted natural antibodies [10]. Marginal Zone (MZ) B cells are another innate-like lymphocyte subset located, as their name implies, at the margins of lymphoid follicles in secondary lymphoid tissues such as the spleen and are a source of low-affinity antibodies that also provide a first line defense against pathogens [11, 12]. These innate-like populations are distinguished from the majority of B cells, often referred to as B-2 B cells, that are effectors of adaptive immunity and constitute most B lymphocytes in secondary lymphoid tissues.

A 2006 study from our laboratory identified phenotypically distinct B-1 and B-2 specified progenitors, thus supporting the conclusion that B cells of the innate and adaptive immune system developed through independent pathways [13]. The ability to resolve these distinct progenitor populations also allowed us to map when they arose during gestation. We found that B-1 B cell progenitors were first detected in the yolk sac at embryonic day (E) 10.5 [14]. Because HSCs in the mouse first arise at E10.5 and few would have been present at that time, it was unlikely that the B-1 progenitors present in the E10.5 yolk sac were derived from stem cells. A study from Yoder and Yoshimoto showing that cells in the yolk sac could generate B-1 but not B-2 B cells well before HSCs emerged was consistent with that conclusion [15]. There is evidence that B-1 B cell progenitors arise in the human fetus as well [16], but this is an area in need of further study. We also found that B-1 progenitors emerged in two additional waves at later times during gestation. However, by three weeks after birth, few B-1 B cell progenitors were present in the bone marrow and that tissue had limited B-1 B cell potential [14, 17]. These mapping studies also identified two distinct waves of B-2 [14] as well as three waves of T cell development [18].

Our analyses also showed that the regulation of fetal and adult B cell development was distinct as exemplified by differences in the requirement for the PU.1 transcription factor. The expression of wild type levels of Spi1, the gene that encodes PU.1, is required for efficient differentiation of adult HSCs into lymphoid specified progenitors [19, 20], and B lymphopoiesis was blocked when it was not expressed [21]. This was also the case in a PU.1 hypomorphic strain generated by deleting an upstream regulatory element (URE) located 14 kb from the Spi1 transcription start site, resulting in an 80% reduction in levels of Spi1 transcripts in bone marrow cells [5, 22]. This strain as originally described is predominantly on a 129Sv background [22], but the mice used in our analyses [5] had been backcrossed onto a C57BL/6 background for seven generations. Because mice with a total Spi1 deletion die in utero or soon after birth [19, 20], we used UREΔ/Δ mice, which appear healthy and survive long-term, to examine the requirement for PU.1 during fetal lymphopoiesis, and this analysis showed that fetal B [14] and T [18] cell development were largely intact. The differential dependence of fetal and adult lymphopoiesis on PU.1 provided yet another example of the differences in the regulation of fetal and adult hematopoiesis that have been described for other lineages. For example, adult tissue resident macrophages emerge in the yolk sac and do not require the Myb transcription factor while macrophage development from adult stem cells is Myb dependent [23].

Our analysis of lymphocyte development in UREΔ/Δ and wild type mice provided no information about the upstream progenitor populations from which the various fetal lymphoid progenitors arose or the precise times during gestation when they did so. However, an elegant lineage tracing study from the Yoshimoto laboratory filled this gap in knowledge [1, 3]. Kobayahsi et al. used an inducible system that labeled hemogenic endothelium and showed that various progenitors with B-1, B-2, and/or T cell potential emerged between embryonic days (E) 7.5-E11.5 independent from HSCs. Their lineage tracing approach identified a B-1 B cell restricted progenitor that arose from hemogenic endothelium [3] thus validating the existence of the phenotypically distinct B-1 B cell restricted progenitor we had defined [13]. Additional studies, including a report from the Carmago laboratory, also documented the fetal emergence of stem cell independent multipotential progenitors (MPPs) in utero [4, 24]. The timing with which the various progenitors emerged in these various waves of fetal development has been discussed in depth, and primary studies and reviews should be consulted for more information [1, 3, 4, 14].

The aforementioned lineage tracing studies [3, 4] concluded that the MPPs that arose in utero independent from HSCs were maintained and generated lymphoid progeny for up to a year after birth [4]. UREΔ/Δ mice also provided a powerful model to investigate the contribution of fetal progenitors to adult lymphopoiesis because, as noted above, fetal lymphocyte development was largely intact while lymphocyte development from adult stem cells was blocked due to their PU.1 deficiency [5, 22]. As a result, it was possible to quantify the total lymphoid output from fetal progenitors, and not just that of lineage marked cells, and evaluate their potential to populate secondary lymphoid tissues with B and T cells. We performed a detailed analysis of the types and numbers of B and T cells present in secondary lymphoid tissues from UREΔ/Δ mice and, as we recently reported [5], the results also showed that fetal lymphoid progenitors were maintained after birth and generated B and T cells that contributed to the adult immune system. However, the issues, as discussed below, were how efficiently and how long they did so.

FETAL LYMPHOCYTE PROGENITORS EFFICIENTLY GENERATE INNATE-LIKE B CELLS

Numerous studies indicated that innate-like B cells, and B-1 B cells in particular, are preferentially generated from fetal progenitors [7, 8, 17]. More recent studies showing that adult stem cells have limited B-1a B cell potential [3, 25] provided further support for this view. In this case, the innate-like B cell compartment should be filled in UREΔ/Δ mice. We quantified CD5⁺ CD23⁻ CD21^−/low sIgM⁺ B-1a and CD5⁻ CD23⁻ CD21^−/low sIgM⁺ B-1b cells in the peritoneal cavity and CD93⁻ CD23⁻ CD21^hi sIgM⁺ MZ B cells in the spleen from UREΔ/Δ mice to test this prediction.

The results showed that B-1 B cells in the peritoneal cavity and MZ B cells in the spleen were present at wild type levels in UREΔ/Δ mice up to one year of age. In fact, the frequency of MZ B cells was higher in UREΔ/Δ compared to wild type mice. These results indicated that the potential of fetal progenitors to fill the adult B-1 and MZ B cell compartments is robust (Figure 1) and that input from adult stem and progenitor cells is not necessary to do so. This conclusion is consistent with a report showing that the MZ and B-1 B cell compartments were intact in mice in which postnatal B cell development was totally blocked by deletion of the recombinase activating gene-2 at birth [26].

Figure 1. Contribution of fetal lymphocyte progenitors and adult stem cells to the immune system.

The figure shows that fetal progenitors can fill the innate-like B-1 and MZ B cell compartments in the adult. However, their contribution to the production of other B and T cell subsets, even in neonates and young adult mice, is limited, and most secondary lymphoid tissues are lymphopenic in the absence of lymphocyte development from adult HSCs. Fetal lymphocyte progenitors are present only transiently after birth, and the figure further shows that by middle age the generation of naive B and T cells is dependent on lymphocyte development from adult HSCs. Nevertheless, the naive B and T cells generated from transient fetal lymphocyte progenitors can mature into activated and memory cells that are maintained long-term. As a result, the activated and memory B and T cell compartments are developmentally heterogeneous and include effectors derived from transient fetal progenitors as well as adult stem cells.

However, several studies have concluded that stem and progenitor cells present in adult bone marrow are a significant source of B-1 B cells as we recently reviewed [2]. For example, Vergani et al. [27] reported, based on lineage tracing of committed B lineage cells in the bone marrow from nineteen-day old mice, that there was extensive postnatal production of B-1a B cells. However, a caveat in the interpretation of the data is that fetal progenitor cells [28] that include B-1 B cell restricted progenitors [13] are present in appreciable numbers in the marrow for several weeks after birth. In contrast, by fifteen weeks after birth the B-1 B cell potential of marrow cells is significantly attenuated [17]. Thus, we consider it likely that the day nineteen marrow cells that were lineage traced by Vergani et al. [27] included fetal derived progenitors with B-1 B cell potential.

If progenitors with B-1 B cell potential are only present during fetal and neonatal development, the question of how the B-1 B cell compartment is maintained long-term after birth arises. In contrast to B-2 B cells, B-1 B cells can self-renew [29]. Thus, even if only a limited number of B-1 B cell progenitors arise in utero and generate a low number of mature B-1 B cells, the latter lymphocytes can expand and fill the B-1 B cell compartment. Our observation that the number of B-1 B cells in murine recipients of a low number of B-1 B cell progenitors increased over time is consistent with this conclusion [13]. Thus, once mature B-1 B cells have been generated, their ability to replicate maintains the size of the B-1 B cell compartment after birth.

FETAL LYMPHOID PROGENITORS MAKE A LIMITED CONTRIBUTION TO POSTNATAL LYMPHOPOIESIS

The first mature B and T cells produced in primary lymphoid tissues are naive, antigen inexperienced cells which then migrate to secondary lymphoid tissues such as the spleen. Because adult lymphopoiesis was blocked in UREΔ/Δ mice, any lymphocytes present after birth must have been derived from fetal progenitors. We enumerated the frequency and total number of naive CD93⁻ CD23⁺ CD21^low sIgM⁺ B-2 cells and CD62L^hiCD44^low naive CD4 and CD8 T cells in secondary lymphoid tissues of young (4-20 week-old) and middle aged (22-46 week-old) UREΔ/Δ mice to determine how efficiently and how long naive B and T cells were generated from fetal progenitors.

One conclusion from this analysis was that the lymphoid potential of fetal progenitors is limited. For example, we found that naive follicular B-2 and CD4 and CD8 T cells were present in the spleen of young UREΔ/Δ mice. However, their number was significantly lower than in the spleen of wild type mice (Figure 2A). This low production of naive lymphocytes was particularly evident upon examination of lymph nodes. Mesenteric lymph nodes (MLNs) in young UREΔ/Δ mice contained few B or T cells (Figure 2B), and peripheral lymph nodes were difficult to detect (Figure 2C). The initial organogenesis of lymph nodes requires lymphoid tissue inducer and initiator cells, but once established naive B and T cells play a role in maintaining their infrastructure [30]. These data indicated that fetal lymphoid progenitors lacked sufficient potential to fill secondary lymphoid tissues. In contrast, the higher number of naive B and T cells in tissues of young wild type mice reflects input from both fetal lymphoid progenitors and adult HSCs (Figure 1).

Figure 2. Secondary lymphoid tissues in UREΔ/Δ mice are acellular.

A. Number of naive follicular B and CD4 and CD8 T cells in the spleen of young and middle-aged wild type (WT) and UREΔ/Δ mice. B. Number of IgM⁺ B and CD4 and CD8 T cells in mesenteric lymph nodes from young and middle-aged WT and UREΔ/Δ mice. C. Inguinal lymph nodes are present in young WT but not UREΔ/Δ mice. Arrows points to the location of the oval inguinal lymph node that is present in WT and absent in UREΔ/Δ mice. The data in panels A and B are based on the detailed data presented in reference [5] from our laboratory.

A second finding was that fetal lymphoid progenitors generated naive B-2 and T lymphocytes for only a limited time after birth, because the latter populations were almost totally absent from the spleen of UREΔ/Δ mice by six months after birth (Figure 2A-B). As Spi1 deletion in mature B [31] and T [32] cells, which do not express detectable PU.1 protein or Spi1 transcripts [33], does not alter their frequency or total number, the lack of naive B and T cells in UREΔ/Δ mice did not occur because they did not express PU.1 at wild type levels. Instead, the fetal lymphoid progenitors either stopped generating naive progeny or, more likely, they were present only transiently after birth and were not maintained long-term. The latter possibility is supported by data showing that no B or T cell progenitors were present in the bone marrow and thymus, respectively, of UREΔ/Δ mice by six months of age [5]. These results indicated that the generation of naive B and T cells past young adulthood is dependent on lymphocyte development from adult HSCs (Figure 1) and that secondary lymphoid tissues are lymphopenic when this does not occur.

Our conclusion that fetal lymphoid progenitors contributed to lymphocyte development for only a limited time after birth is not in agreement with one lineage tracing study which concluded that they did so for a year or more [4]. We cannot exclude the possibility that stem cell independent fetal progenitors capable of long-term survival and differentiation after birth may not have emerged in UREΔ/Δ mice because they were PU.1 dependent. We do not think this is the case, because our extensive analysis of fetal lymphocyte development in UREΔ/Δ mice only documented the absence of the yolk sac wave of B-1 B cell development [14] and, as described in the preceding section, this did not compromise the size of the adult B-1 B cell compartment. Instead, our data are more in line with the Kobayashi et al. study [3] reporting that the number of fetal derived, stem cell independent MPPs had declined by ~75% by four months after birth [1].

ACTIVATED/MEMORY CELLS DERIVED FROM FETAL PROGENITORS ARE MAINTAINED LONG-TERM

Despite the fact that fetal progenitors produced a low number of naive B and T cells for a limited time after birth, UREΔ/Δ mice survived long-term and exhibited no signs of infection, suggesting they had a functional immune system. This conclusion was supported by 16S rRNA sequencing of fecal samples which revealed that the microbiome in UREΔ/Δ and wild type mice was similar [5]. Thus, the naive B and T cells that emerged from transient fetal progenitors must have matured into effector cells that provided immune protection. We tested this hypothesis by performing detailed phenotyping to identify activated and memory B and T cells in UREΔ/Δ mice up to one year of age. This approach was critical, because characterization of B and T cells in adult tissues with IgM, CD4, or CD8 alone would not distinguish naive from antigen experienced cells. The results allowed us to conclude that some naive B and T cells that emerged from fetal progenitors became antigen activated and matured into effector cells. For example, the spleen of young UREΔ/Δ mice contained CD4 and CD8 central and effector memory T cells as well as IgA⁺ and CD80⁺ activated B cells. These populations were also present in the small and large intestines [5].

We also found that these antigen experienced cells were maintained in secondary lymphoid tissues through at least middle age. For example, the spleen of UREΔ/Δ mice contained sIgM⁺ CD80⁺ activated B cells and CD8⁺ T central memory and effector memory cells, plasma cells/plasmablasts; CD4 and CD8 T central and T effector memory cells were also present in the intestines of UREΔ/Δ mice up to one year of age. Furthermore, their total number was often the same as observed in wild type mice. For example, the number of sIgM⁺ CD80⁺ activated B cells in the spleen of middle aged UREΔ/Δ mice was the same or slightly higher compared to wild type mice. Similarly, the total number of CD4 and CD8 T effector memory cells was similar in the small and large intestines of UREΔ/Δ and wild type mice [5]. These data suggest that the antigen experienced B and T cells present in adult secondary lymphoid tissues are not derived exclusively from adult stem cells.

It will be of interest to determine if the various immune effectors in adult UREΔ/Δ mice included previously described fetal-derived populations. In particular, the Rudd laboratory identified a hierarchy of CD8⁺ T cells that included neonatal and adult subpopulations. The neonatal CD8 T cells exhibited robust and diverse cytokine secretion and rapid proliferation [34, 35]. It is also interesting that these neonatal T cells, which are derived from fetal lymphoid progenitors, provided protection against bacterial infection, because mice in which they were absent had increased bacteria in the spleen [36]. This is relevant in view of our data showing a relatively normal microbiome in UREΔ/Δ mice.

Another key question was whether B-1 B cells, regardless of when during ontogeny they were generated, colonized the intestines and contributed to gut immunity. There are multiple, independent reports indicating that they do not do so [37-41] as well as papers indicating that they make a major contribution to intestinal immunity [27, 42-45]. The analysis of UREΔ/Δ mice showed that IgA expressing cells that could have been B-1 B cell derived were present in the small and large intestines [5]. However, we could not optimize a reliable phenotyping protocol to determine if the B cells we identified were bona fide B-1 B cells. Unequivocal identification of B-1 B cells in tissues other than serous cavities is a challenge, and the fact that activation could result in alteration of cell surface markers added to the difficulty in identifying cells of B-1 origin. This is even more of a challenge in humans because there is no widely agreed upon phenotype for B-1 B cells.

CONCLUSIONS

It is now evident that various progenitor populations with lymphoid potential arise in multiple waves of fetal development that are often stem cell independent. There is also consensus that fetal derived lymphoid progenitors are maintained after birth and generate lymphoid progeny. However, the analysis of UREΔ/Δ mice indicated that their lymphoid output and longevity after birth are limited. As a result, secondary lymphoid tissues were generally lymphopenic in the absence of adult lymphopoiesis. We thus conclude that lymphocyte development from adult HSCs is necessary to fill secondary lymphoid tissues and that fetal progenitors lack the potential to do so. Thus, fetal lymphoid progenitors cannot substitute for adult HSCs in the generation and maintenance of the full complement of B and T cells that constitute the adult immune system.

We nevertheless propose that they make two obligate contributions to post-natal immunity. One is to establish the innate-like B cell compartment. In fact, we suggest that the majority of B-1 and MZ B cells in adult lymphoid tissues are derived from transient fetal lymphoid progenitors. The fact that the B-1 and MZ B cell compartment were filled at wild type levels in UREΔ/Δ mice supported this conclusion. A second major role is to generate sufficient numbers of B and T cells so that the newborn and neonate have the capacity to respond to antigenic challenge until lymphopoiesis from adult HSCs is established. This protection is particularly important in the intestines which are colonized by bacteria soon after birth [46]. In view of this, we suggest that the multiple types of B cells, including a subset of regulatory B cells that produce antibodies with a wide range of microbiota reactivity [47], and T cells present in the human fetal intestines [48-51] are the progeny of transient fetal lymphocyte progenitors.

A final consideration is that the B and T cells generated from fetal progenitors may be distinct from those that are the progeny of adult HSCs as discussed in the preceding section with regards to CD8 T cells. They may also have distinct tissue homing properties. For example, while fetal derived B and T cells were present in spleen and intestines, few were present in lymph nodes (Figure 2B). We also reported that that few if any plasma cells/plasmablasts were present in the bone marrow of UREΔ/Δ mice regardless of age [5]. It will be interesting to determine if this is the case after vaccination; if plasma cells/plasmablasts are still not observed in the marrow, it would provide further evidence that the types of B cells generated from fetal progenitors are distinct from those that are the progeny of adult HSCs. This discussion emphasizes the need to compare the properties of B and T cells generated from fetal progenitors versus adult HSCs in more detail.

Highlights.

• Lymphocyte progenitors emerge in multiple stem cell dependent and independent fetal waves of development

• Fetal derived lymphocyte progenitors generate B and T cells for a limited time after birth

• Fetal derived progenitors have the potential to fill the innate-like B lymphocyte compartments

• Activated and memory B and T cells generated from fetal lymphocyte progenitors are maintained long-term after birth