Prenatal development of human immunity (original) (raw)

Science. Author manuscript; available in PMC 2022 Jun 25.

Published in final edited form as:

PMCID: PMC7612900

EMSID: EMS145613

Jong-Eun Park,#1 Laura Jardine,#2 Berthold Gottgens,3,4 Sarah A. Teichmann,1,5,† and Muzlifah Haniffa1,2,6,†

Jong-Eun Park

1Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridge CB10 1SA, UK

Laura Jardine

2Biosciences Institute, Newcastle University, Faculty of Medical Sciences, Newcastle upon Tyne NE2 4HH, UK

Berthold Gottgens

3Department of Haematology, University of Cambridge, Cambridge CB2 2XY, UK

4Wellcome and MRC Cambridge Stem Cell Institute, University of Cambridge, Cambridge CB2 2XY, UK

Sarah A. Teichmann

1Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridge CB10 1SA, UK

5Department of Physics/Cavendish Laboratory, University of Cambridge, JJ Thomson Ave., Cambridge CB3 0HE, UK

Muzlifah Haniffa

1Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridge CB10 1SA, UK

2Biosciences Institute, Newcastle University, Faculty of Medical Sciences, Newcastle upon Tyne NE2 4HH, UK

6Department of Dermatology and NIHR Newcastle Biomedical Research Centre, Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne NE2 4LP, UK

1Wellcome Sanger Institute, Wellcome Genome Campus, Hinxton, Cambridge CB10 1SA, UK

2Biosciences Institute, Newcastle University, Faculty of Medical Sciences, Newcastle upon Tyne NE2 4HH, UK

3Department of Haematology, University of Cambridge, Cambridge CB2 2XY, UK

4Wellcome and MRC Cambridge Stem Cell Institute, University of Cambridge, Cambridge CB2 2XY, UK

5Department of Physics/Cavendish Laboratory, University of Cambridge, JJ Thomson Ave., Cambridge CB3 0HE, UK

6Department of Dermatology and NIHR Newcastle Biomedical Research Centre, Newcastle Hospitals NHS Foundation Trust, Newcastle upon Tyne NE2 4LP, UK

#Contributed equally.

exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works

Abstract

The blood and immune systems develop in parallel during early prenatal life. Waves of hematopoiesis separated in anatomical space and time give rise to circulating and tissue-resident immune cells. Previous observations have relied on animal models, which differ from humans in both their developmental timeline and exposure to microorganisms. Decoding the composition of the human immune system is now tractable using single-cell multi-omics approaches. Large-scale single-cell genomics, imaging technologies, and the Human Cell Atlas initiative have together enabled a systems-level mapping of the developing human immune system and its emergent properties. Although the precise roles of specific immune cells during development require further investigation, the system as a whole displays malleable and responsive properties according to developmental need and environmental challenge.

Animal model systems have provided fundamental evidence that shapes our understanding of developmental hematopoiesis. Studies performed in mouse, zebrafish, and chicken have established that blood and immune system development occur across distinct anatomical sites (Fig. 1). The first blood cells are extraembryonic, developing in close association with endothelial cells of the yolk sac (1). Embryonic hematopoietic stem cells (HSCs), capable of repopulating an adult host in transplant assays, originate from the aorta gonad mesonephros (AGM) region (2). The fetal liver and bone marrow (BM) are subsequently seeded by both yolk sac-derived progenitors and AGM-derived HSCs (3). However, developmental timelines are not chronologically identical between species. For example, mouse fetal thymus is notably immature compared with human thymus, which supports complete naïve T cell differentiation in utero (4). Furthermore, some population-defining markers are poorly conserved, making it difficult to directly apply findings from animal studies to humans. Increasingly, the influence of maternofetal microbial exposure on the fetal immune development is recognized, and both commensal and pathogenic microbial repertoire differs among species (see accompanying reviews). Studies directed at human immune development have been hampered by tissue access and experimental limitations, but single-cell multi-omics technologies have expedited new findings. In this review, we discuss how these technologies have provided an unprecedented view of early life immunity. We describe key insights into how immune development is layered across time and space and explain how immune cells both prepare the fetus for antigen challenge and adopt noncanonical roles in development.

Temporal and spatial development of the human immune system.

The development of blood and immune systems during early human life occurs over several anatomical sites. The major site of hematopoiesis changes from the extraembryonic yolk sac to the intraembryonic AGM, liver, and BM. T cell differentiation and maturation are confined to the thymus. Immune cells seed other lymphoid or peripheral organs—including lymph nodes, skin, intestine, kidney, and lung—and adapt to the respective organ environment. Diverse immune cell types develop and mature at different gestational stages, which is necessary to establish tolerance and functional response based on developmental needs. This prepares the developing embryo and fetus for antigen exposure during pregnancy and after birth. ILCP, ILC precursor; CDR3, complementary-determining region 3; TdT, terminal deoxynucleotidyl transferase.

From single cells to system-level development

The challenge of unraveling blood and immune system development in the prenatal human requires a high-performance tool kit. Single-cell RNA sequencing (scRNA-seq) has emerged as a powerful tool for the systematic understanding of the immune system, permitting an unbiased identification of cell state and the resolution of complex mixtures of cells (5). Droplet-based scRNA-seq methods, such as 10X, are now scalable to the extent that whole organs can be adequately sampled. For example, our group has profiled single cells from yolk sac and liver to reconstruct early hematopoiesis, from thymus to capture T cell development, and from skin and kidney to elucidate the seeding of peripheral organs (4, 6). Computational techniques have permitted comparison of cell states across tissues and prediction of critical receptor ligand interactions that shape immune cell fate in specific tissues (7). Correlation with imaging techniques—for example, in situ transcriptomics—has allowed comprehensive characterization of tissue micro-environments (4, 7–9). Developmental trajectories have been inferred within single tissues, as cells are captured at varying stages of differentiation and by integrating samples from a range of gestational ages. This tool kit is beginning to provide a comprehensive overview of early immune development. Meanwhile, considerable challenges remain in tracing the origin of specific immune cells to distinct waves of hematopoiesis. Advances in single-cell DNA sequencing combined with analytical techniques to track distinct clones may bring us closer to this goal.

The developing immune system in space and time

In this section, we follow human immune system development across space and time. We begin by discussing cell types as they first emerge in the yolk sac or fetal liver, before considering the thymus as a key site of T cell development. This cannot be an exhaustive description of immune composition because about 40 immune cell states have been identified in these tissues to date. Instead, we focus on how single-cell multi-omics approaches have advanced our understanding of the human fetal immune system (Fig. 2).

Overview of single-cell studies detailing the developing human immune system.

Diverse single-cell methods (depicted in color) have been applied to generate a comprehensive atlas of human immune system development. In many studies, multiple organs have been sampled together to investigate the migration, adaptation, and compartmentalization of immune cells. (Studies are indicated by the reference number, and dotted lines link the different organs sampled in each study.)

Yolk sac and AGM

An analysis of the human embryonic yolk sac demonstrates the presence of HSC-like progenitors, macrophages, mast cells (MCs), natural killer (NK) cell progenitors, and innate lymphoid cell (ILC) progenitors alongside megakaryocytes and erythroid cells from four postconception weeks (PCW) (6).

Macrophage origin has been intensively studied because tissue macrophages arise independently from HSCs and self-renew under homeostatic conditions in mouse models (10). Tissue-resident macrophages in the liver, lung, brain, and epidermis were shown by fate mapping to arise from yolk sac hematopoiesis through the erythromyeloid progenitor (11, 12). Although the yolk sac contribution is retained in some tissues (e.g., the liver, brain, and epidermis), macrophages are gradually replaced by HSC-derived monocytes at other sites (e.g., the gut, lung, and heart). This process depends in part on how "open" the niche remains to circulating cells (10). In the mouse, the precise contributions of the first versus second waves of yolk sac hematopoiesis and whether the macrophages arise from a monocyte intermediate remain unresolved (10). In human fetal development, tissue-specific macrophages are observed from the earliest time points sampled (6, 13, 14). Single-cell dissection of human AGM revealed a distinct hemogenic endothelial population that gives rise to macrophages (13). By 6 PCW, the embryonic pancreas is laden with macrophages, microglia accompany the developing brain, and Hofbauer cells line the placenta (7, 14, 15). Identification of these cells in appreciable numbers before the onset of fetal liver hematopoiesis at 6 to 9 PCW lends support to yolk sac or AGM-derived macrophages seeding peripheral tissues. Attempts to use transcriptional similarity between yolk sac macrophages and fetal liver macrophages to parse tissue macrophage ontogeny are not sufficiently reliable owing to environmentally related gene expression after tissue residency. However, these profiles have allowed characterization of macrophage diversity essential for development, for example, erythroid island macrophages providing support for erythropoiesis and Kupffer cells with prominent scavenging function in the fetal liver (6).

In parallel to macrophages, fate-mapping studies have demonstrated that tissue MCs arise from both yolk sac and HSC-derived precursors in mouse and that patterns of yolk sac MC retention are tissue specific (16, 17). In human development, a clear MC signature is present in both the yolk sac and fetal liver (6). Connective tissue MCs in fetal skin and kidney are closely related to fetal liver MCs by single-cell gene expression profile (6). This early dedication by the embryo to MC production is puzzling. The best-characterized function of MCs is their participation in allergic responses on immunoglobulin E (IgE) binding via the high-affinity IgE receptor (18). Liver and yolk sac MCs appear ill prepared for this task, because neither expresses the IgE receptor alpha subunit gene (FCER1A) (6). Early MC production may occur to equip developing mucosal sites and connective tissues with resident immune cells or to provide a pool of pathogen-associated molecular pattern-responsive innate effectors. However, additional functions in supporting angiogenesis are predicted. In mice, embryonic skin MCs express genes involved in vascular and neural patterning (6, 16). In adult mammals, MCs support both physiological and inflammatory angiogenesis (18). The role of MCs in prenatal vascular development warrants further investigation.

NK cells, ILC progenitors, and their common lymphoid progenitors can be identified from yolk sac and fetal liver single-cell transcriptome (6, 19). In later stages, they are found as more diverse and differentiated cells in multiple fetal organs (9, 20). In contrast to the maternal decidual NK cells whose role during pregnancy has been well characterized (7, 21), our understanding of fetal NK cell function to date is limited. Although fetal NK cells are considered to be immature and hyporeactive compared with adult NK cells, they already possess killer activity (22, 23). Moreover, fetal or infant NK cells resemble their adult counterparts at several levels, suggesting that they are poised to respond when the right stimuli, such as viral infections, are present (23). Concordantly, NK cells are abundant in infant intestines, are equipped with cytolytic granules, and display superior degranulation activity compared with adult intestinal NK cells (20). In addition to NK cells, other ILCs have been shown to be enriched in the fetus compared with infants (24). Among them, innate lymphoid tissue inducer (LTi) cells play a critical role in the formation of secondary lymphoid organs (25, 26). By interacting with stromal cells, LTi cells induce positive feedback to recruit additional LTi cells as well as other immune cells, generating a lymphoid environment (27). Thus, innate lymphocytes develop very early in the human embryo and are involved in both tissue protection and remodeling.

This earliest wave of hematopoiesis in the yolk sac displays dedication to immune cells with structural and physiological roles alongside equipping the embryo with a basic repertoire of innate immune effectors. The precise roles of these cells in tissue development and the checkpoints that prevent damaging immune responses in utero require further investigation.

Liver and BM

Definitive HSCs can generate the full complement of erythroid, megakaryocyte, myeloid, and lymphoid cell lineages in fetal liver, but neutrophils remain absent until BM hematopoiesis is established (28).

In contrast to macrophages, monocytes and dendritic cells (DCs) are considered HSC-dependent populations. In the mouse, both are traceable to a clonogenic precursor in BM named the macrophage-DC progenitor (29). In human development, the first signs of DC production are seen in the fetal liver from around 6 PCW (6). Conventional DC1, DC2, and plasmacytoid DCs are found in fetal tissues—including the lung, spleen, skin, and thymus—from 12 PCW and are relatively abundant compared with adult tissue DCs (30). Fetal DCs, like their adult counterparts, are capable of migrating, responding to Toll-like receptor ligation, and stimulating T cell proliferation and activation (30). Fetal DCs have the particular capacity to induce regulatory T cell differentiation, promote T cell interleukin-4 production, and inhibit T cell tumor necrosis factor–α (TNFα) production via arginase 2 (30). Thus, DCs play an important role in maintaining tolerance during fetal life.

The B cell lineage is first observed in the fetal liver from 7 PCW in the form of B cell precursors; mature B cells are present only after 9 PCW (6). This has been attributed partly to the change in HSC-intrinsic potential to generate B cells and the liver microenvironment support for B cell differentiation (6). At mid-gestation, the BM becomes the major source of B cells, and mature B cells are abundantly enriched in spleen (31). Although fetal B cells achieve diverse repertoire from early stages (24, 32), the formation of germinal centers is attenuated until antigen exposure after birth, which is accompanied by active somatic hypermutation (33). Comparing intestinal B cells from second-trimester fetuses to infants with single-cell mass cytometry combined with B cell receptor repertoire analysis nicely demonstrated that fetal intestinal B cells are primarily follicular and transitional B cells, whereas plasma B cells are enriched in infants (24).

Another interesting aspect of B cell development that has been intensively studied in the mouse model is the tiered development of innate-like B-1 cells, which predominate in early gestation and are followed by conventional B-2 cells (34). However, the precise identity of human B-1-like cells has not yet been resolved (35). Future studies to generate a single-cell atlas of human fetal BM and spleen will provide a better view on human B cell ontology, highlighting organ-specific differences in the niche factors that support B cell differentiation.

“The clinical implications of fetal immune development and function reach far beyond life in utero.”

Thymus and peripheral organs

The thymus provides an environment essential for T cell development. Early lymphoid progenitors originating from the fetal liver migrate into the thymus at 8 PCW, where they develop into naïve T cells (36).

Development and maturation of the thymus are mediated by an interplay between thymic stromal cells and immune compartments, which has been largely studied in mouse models. Comprehensive single-cell transcriptome profiling of cellular constituents of developing human thymus showed extensive communication between thymic epithelial cells, mesenchymal cells, early thymic progenitors, developing and mature T cells, and other immune cells (4, 19). The proportion of each cell population also shows coordinated change across development, further proving the importance of harmony between multiple cell types for organ maturation (4).

Single-cell studies on fetal liver and thymus revealed detailed molecular signatures accounting for the transition from early thymic progenitors into naïve T cells (4, 6, 19). Hu and colleagues focused on the molecular profile of thymus-seeding progenitors (19). Our group extended this analysis toward later stages in development (4). Together, these findings revealed a continuous trajectory from early thymic progenitors developing into multiple mature T cell types.

Naïve T cells egress from the thymus and migrate into other tissues. Circulating T cells are observed at 10 to 11 PCW after functional thymic development (37). The absence or presence of microorganisms in the fetal environment remains a matter of debate (see accompanying reviews). Although a healthy pregnancy is most likely sterile, noninherited maternal alloantigens and microbial by-products may potentially activate the fetal immune system. To avoid damaging alloreactivity, the fetus needs to maintain tolerogenic immunity. Consequently, naïve T cells generated from the fetus are more likely to acquire a regulatory T cell fate compared with adult naïve T cells (38). Fetal regulatory T cells suppress the proliferation and cytokine secretion of other fetal T cells that are potentially self-reactive (39).

Memory T cells have been identified in the fetal intestine, highlighting the potential of fetal T cells to respond to foreign antigens (9, 24, 40, 41). Studies on intestinal CD4+ T cells by single-cell techniques combined with repertoire sequencing identified the existence of memory T cell populations and regulatory T cells with the signature of clonal expansion, highlighting the balance between activation and suppression of adaptive immune response in the fetus (24, 42). Intestinal CD4+ T cells can also play a role in promoting development, as shown for the case of moderate TNFα expression (41). Thus, fetal adaptive immunity is substantially more mature than previously expected. Active areas of future research on the fetal immune system include the antigenic cues underlying fetal T cell activation and the roles they play in fetal development and protection.

Through this snapshot of fetal immune development across time and space, we note the emergence of both innate and adaptive immune cells with distinctive properties compared with their adult counterparts. Among the components missing from this overview are neutrophils. Current evidence suggests that around one-third of fetal BM cells are neutrophils or their precursors at 10 to 13 PCW, increasing to two-thirds at 21 PCW (43). Infants born prematurely or small for gestational age have lower circulating neutrophil counts, lower neutrophil reserve, and higher mortality from sepsis (44). Understanding how the fetal neutrophil compartment operates will provide insights into how early-life immune defense can be supported.

Conclusion

Multi-omics suspension and spatial-based technologies have provided ideal platforms to dissect and reconstruct the developing immune system (4, 6, 9, 13, 19, 24, 41, 42). Many areas of uncertainty remain to be unraveled (Fig. 3). How do hematopoietic progenitors change throughout development? How do different tissue niches such as yolk sac, liver, BM, thymus, and spleen affect the progenitor populations and developing immune cells? How do immune cells migrate to and adapt in peripheral nonlymphoid tissues? How does the immune system communicate, learn, and form memory for future encounters?

Key questions to be addressed in future studies of immune system development.

The diagram depicts pertinent questions relating to the developing immune system. How do the HSCs change in their potential throughout development? How do diverse hematopoietic niches differ from each other? What determines the migration of immune cells to the target organs, and how do they adapt to a new tissue environment? Single-cell profiling and spatial profiling techniques are now providing answers to these questions by assessing the immune system as a whole and identifying emergent properties of the collective.

Completion of the developing immune atlas by focusing on the organs and time points that are currently missing, extending the analyses for comparison with the adult immune cells, and system and cross-species comparisons will provide further knowledge about how the human immune system evolved and is established and sustained. The clinical implications of fetal immune development and function reach far beyond life in utero. Fetal-specific hematopoietic progenitor cells are now recognized as likely cells of origin for blood cancers, including Down syndrome–associated acute megakaryoblastic leukemia, juvenile myelomonocytic leukemia, and infant acute lymphoblastic leukemia. Early-onset primary immunodeficiencies with impaired response to pathogen challenge and/or autoimmunity may also be influenced by developmental cues and changes. In these settings, aberrant hematopoiesis also results in abnormal immune function. The biological insights from in-depth understanding of the developing immune system promise to revolutionize stem cell transplantation and tissue engineering for immunotherapy and regenerative medicine in the near future.

Acknowledgments

We thank J. Eliasova for graphical images.

Funding

We acknowledge funding from the Wellcome Human Cell Atlas Strategic Science Support (WT211276/Z/18/Z). M.H. is funded by Wellcome (WT107931/Z/15/Z), the Lister Institute for Preventive Medicine, and the NIHR Newcastle Biomedical Research Centre. S.A.T. is funded by Wellcome (WT206194), ERC Consolidator and EU MRG-GRammar awards, and the Chan Zuckerberg Initiative (CZF2019-002445). B.G. is a Wellcome Investigator (206328/Z/17/Z) and is also supported by core funding from Wellcome and MRC to the Cambridge Stem Cell Institute. L.J. is funded by an NIHR Academic Clinical Lectureship. J.-E.P. is supported by an EMBO Advanced Fellowship (ALTF 623-2019).

Footnotes

Competing interests: In the past 3 years, S.A.T. has consulted for Biogen, GenenTech, and Roche and is a member of the ForeSite Labs Scientific Advisory Board.

References and Notes

1. His W. Lecithoblast und angioblast der wirbelthiere. Histogenetische studien. Vol. 26 BG Teubner; 1900. [Google Scholar]

4. Park J-E, et al. Science. 2020;367:eaay3224. [Google Scholar]

5. Efremova M, Vento-Tormo R, Park J-E, Teichmann SA, James KR. Annu Rev Immunol. 2020;38 annurev-immunol-090419-020340. [PubMed] [Google Scholar]

11. Perdiguero E Gomez, et al. Nature. 2015;518:547–551. [Google Scholar]

14. Banaei-Bouchareb L, Peuchmaur M, Czernichow P, Polak M. J Endocrinol. 2006;188:467–480. [PubMed] [Google Scholar]

27. van de Pavert SA, Mebius RE. Nat Rev Immunol. 2010;10:664–674. [PubMed] [Google Scholar]

39. Michaëlsson J, Mold JE, McCune JM, Nixon DF. J Immunol. 2006;176:5741–5748. [PubMed] [Google Scholar]

41. Schreurs RRCE, et al. Immunity. 2019;50:462–476.:e8. [PubMed] [Google Scholar]

43. Kelemen E, Calvo W, Fliedner TM. Atlas of Human Hemopoietic Development. Springer; 2013. [Google Scholar]