Fetal Hematopoietic Stem Cells Are the Canaries in the Coal Mine That Portend Later Life Immune Deficiency

Abstract

Disorders of the blood system are a significant and growing global health concern and include a spectrum of diseases ranging from aplastic anemia and leukemias to immune suppression. This array of hematological disorders is attributed to the fact that the blood system undergoes a perpetual cycle of turn over with aged and exhausted red and white blood cells undergoing daily replacement. The foundational cells of this replenishment process are comprised of rare hematopoietic stem cells (HSCs) located in the bone marrow that possess the dual function of long-term self-renewal and multilineage differentiation. This constant turnover makes the hematopoietic system uniquely vulnerable to changes in the environment that impact multilineage differentiation, self-renewal, or both. Notably, environmental endocrine-disrupting exposures occurring during development, when HSCs are first emerging, can lead to alterations in HSC programming that impacts the blood and immune systems throughout life. In this review, we describe the process of fetal hematopoiesis and provide an overview of the intrauterine environmental and endocrine-disrupting compounds that disrupt this process. Finally, we describe research opportunities for fetal HSCs as potential sentinels of later-life blood and immune system disorders.

Progress in the field of prenatal programming has identified epigenetic changes to the chromatin in the absence of DNA mutations as the key mechanism supporting the hypothesis that the preconception and intrauterine environment impacts health and disease susceptibility throughout the life course (1, 2). Referred to as the developmental origins of health and disease (DOHaD), this hypothesis originally was used to describe the connection between nutritional deprivation during pregnancy and cardiovascular disease risk as adults (3). A number of studies have expanded DOHaD to include a spectrum of disorders ranging from neurobehavioral deficits, obesity, reproductive disorders, allergies and asthma (4). Of these, developmental exposures that lead to later life immune suppression are among the least well studied but are no less profound (5).

In this review, we will summarize the general process of hematopoiesis, including the unique aspects of blood formation occurring in the fetus. We will then survey the effects of several classes of endocrine-disrupting chemicals on fetal hematopoiesis followed by an analysis of the future directions of the field and suggest concepts that should be considered to address the critical data gaps linking effects on development of the hematopoietic system to later-life blood and immune system disorders.

Comparison of Hematopoiesis Between the Adult and Fetus

Hematopoiesis occurs in the bone marrow throughout the postnatal life with long-term self-renewing cells residing in a hypoxic environment located in the endosteum region of the bone marrow (6, 7). Self-renewal is the process whereby cell division of a stem cell produces two daughter cells. One cell retains the multipotency of the parent, defined as complete blood system reconstitution potential, whereas the other daughter cell begins a process of multilineage differentiation as illustrated by Figure 1. This dual function of long-term self-renewal and multilineage differentiation potential is facilitated in part by the differential energetic requirements of quiescent resting vs activated hematopoietic stem cells (HSCs). Regulation of the energetic state is governed both by intrinsic regulatory factors active in the cell and extrinsic signals provided by the hematopoietic microenvironment or niche. Notably, the endosteum region of the bone marrow is hypoxic, and HSCs stationed in this niche produce ATP anaerobically by glycolysis (8). Regulated by the oxygen sensing transcription factor hypoxia inducible factor alpha and its dimerization partner Aryl hydrocarbon receptor nuclear translocator/hypoxia inducible factor beta, enforcement of a glycolytic state maintains a low oxidative state protecting the DNA of long-term self-renewing cells from damage (9–12). In comparison, transition out of long-term self-renewal is characterized by migration to the normoxic vascular region of the bone marrow and is accompanied by a switch to aerobic respiration. These cells may continue to maintain short-term self-renewal properties; however, elevated reactive oxygen species produced during oxidative phosphorylation signal the initiation of multilineage differentiation (13, 14).

Figure 1.

Hematopoietic processes impacted by endocrine-disrupting compounds. The dual process of long-term self-renewal and multilineage differentiation is a highly ordered and regulated hierarchical process illustrated here. Functional characterization of hematopoietic stem and progenitor cells is facilitated by cell sorting of these cells based on the differential expression of surface proteins identified on the left side of the diagram. Proteins expressed on human hematopoietic stem and progenitor cells are colored in blue, and murine markers are in black. In general, hematopoiesis is characterized by HSCs giving rise to multipotent and subsequently oligopotent progenitors concomitant with a loss of long-term self-renewal potential. Oligopotent progenitors consist of common lymphoid precursors (CLPs), common myeloid precursors (CMPs), myeloid/erythrocyte precursors (MEPs), and granulocyte/macrophage precursors (GMPs). In response to intrinsic and extrinsic factors present in the hematopoietic niche, oligopotent progenitors become further specified to become lineage restricted progenitors. Linage-restricted progenitors consist of megakaryocyte progenitors (MkPs), erythrocyte progenitors (EPs), granulocyte progenitors (GPs), promacrophages (ProMs), prodendritic cells (ProDCs), pro-B, pro-NK, and pro-T cells. Lineage-restricted progenitors undergo additional maturation, selection, and specification to become mature effector cells. A nonexhaustive list of environmental contaminants and endocrine-disrupting compounds known to impact fetal hematopoiesis and described in this review are identified in red, along with the stage(s) of hematopoiesis affected by the specific agent. Additionally, endocrine-disrupting compounds suspected of impacting hematopoiesis based on their known mode(s) of action are listed in green.

In comparison with the postnatal life, hematopoiesis in the fetus is more complex, because the anatomical site and state of energy production is highly dynamic and is dependent on the stage of gestational development and anatomical location. Specifically, in mice and humans, hematopoiesis begins in the yolk sac at gestational day 7.5 and 30 days after conception respectively (15). Yolk sac blood islands in both mice and humans give rise to a population of nucleated erythrocytes and macrophage-like cells; however, there is debate about the contribution of these yolk sac HSCs to subsequent hematopoietic compartments in the fetus and adult (16–18).

Subsequent to primitive hematopoiesis occurring in yolk sac, an endothelial precursor located in the aorta-gonad-mesonephros is responsible for producing the first definitive HSCs in the fetus (19, 20). Definitive HSCs are so named by virtue of their capacity to reconstitute lethally irradiated mice, thus demonstrating long-term self-renewal and multilineage differentiation potential (21–23). Nearly simultaneous to their emergence in the aorta-gonad-mesonephros is the appearance of definitive HSCs in the placenta. Remarkably, the identification of a HSC niche within the placenta in both mice and humans has only emerged within the last decade (23–25). In both species, the placenta provides a microenvironment that promotes proliferation of HSCs while preventing multilineage differentiation (24–26). This proliferation is followed by migration to the fetal liver where both lymphoid and myeloid lineage cell production increases to seed the peripheral immune organs. Finally, fetal HSCs migrate to the bone marrow where they will reside for the remainder of one's life. These anatomical and developmental transitions, are accompanied by energetic changes such that ATP production alternates between glycolysis and oxidative phosphorylation depending on the developmental stage of the fetus (19, 27). Given that cellular metabolic processes have recently emerged as critical for the maintenance of HSCs (9, 28, 29), it raises the potential that exogenous factors present in the intrauterine environment that interfere with energy regulation, migration, cell cycle, or differentiation may have significant impacts on the long-term programming and function of HSCs. For example, functional HSCs, present in the placenta and cord blood and often discarded, represent a highly sensitive cell population that could be functionally interrogated at birth as a potential sentinel for later-life blood system and immune health.

Methodologies for Determining HSC Fitness

A key question raised by the DOHaD hypothesis is how do we identify at the time of birth an individual vulnerability to later-life adverse health maladies. In the case of the blood and immune system, assessing the fitness of hematopoietic progenitor and stem cells holds promise for the identification of characteristics that may give rise to diseased cells later in life. Notably, environmental health scientists continue to use both in vitro and in vivo tools to provide a detailed picture of the diverse mechanisms by which toxicants, pharmaceuticals, and endocrine-disrupting compounds impact hematopoiesis (30, 31). Although a number of these functional assays have been developed in rodent models, most are readily adaptable to humans and if combined with rigorous exposure assessment, have potential to identify an individual's or population's risk of later-life hematopoietic disease. In the next section, we will review the effects that different classes of environmental agents with endocrine-disrupting properties have on hematopoiesis, focusing on developmental exposures where data are available. We will cover dioxins and polychlorinated biphenyls, cigarette smoking, sex hormone receptor-dependent regulation and pesticides.

Dioxins and Polychlorinated Biphenyls

The immunotoxicity of dioxins have been known for more than 40 years following studies that demonstrated 2,3,7,8 Tetrachlorodibenzo-p-dioxin (TCDD) causes thymic atrophy in mice, rats, and guinea pigs (32). Subsequent studies in rodent models demonstrated that TCDD is myelotoxic to the bone marrow (33). The first demonstration that prenatal TCDD exposure impacts the hematopoietic system came from maternal exposure studies in pregnant mice demonstrating that exposure to TCDD inhibits prothymocyte activity in the fetal liver and fetal bone marrow of the developmentally exposed fetuses (34). Although TCDD is an endocrine disruptor because of its antagonism of estrogen receptor-α signaling (35), the generation of aryl hydrocarbon receptor (AHR)-deficient mice demonstrated that all the immune and hematopoietic toxicity of TCDD is mediated by the AHR (36, 37). More recently, the knowledge that TCDD impairs hematopoiesis through its activation of the AHR has led to a more fundamental understanding of the AHR as an important regulator of hematopoiesis, quiescence, and long-term self-renewal (38–42). Despite growing recognition on the importance of AHR-dependent signal transduction for adult hematopoiesis, far less is understood about the specific role for the AHR in the development of the fetal hematopoietic system. To address this gap, our laboratory has demonstrated that prenatal TCDD exposure attenuates HSC differentiation into lymphocytes (43), and we have further observed that developmental TCDD exposure impairs long-term self-renewal (Laiosa, M.D., unpublished data). Additional studies demonstrated that developmental AHR activation leads to elevated autoimmunity and also impairs host response to influenza later in life (44–48). In the case of influenza, the mechanism of this later life immune suppression appears to be through changes in DNA methylation with an initial wave of differential methylation induced by TCDD-induced AHR during development. A second wave of differential methylation of the DNA affecting far more genetic loci occurs in adulthood after the influenza infection (49). This elegant study illustrates mechanistically a 2-hit model whereby a prenatal exposure leads to epigenetic modifications of a progenitor cell population that then has significant impacts on the function of the progeny of these cells during a later-life secondary environmental insult or disease state.

Although the literature is far less extensive on the effects of polychlorinated biphenyls on hematopoiesis, several Polychlorinated biphenyl (PCB) congeners have been found to be developmentally immunosuppressive, suggestive of an effect on a hematopoietic progenitor. In particular, attenuated T cell-dependent responses in Atlantic Gray seal pups were associated with elevated dioxin-like serum PCBs (50). In people, seminal epidemiological studies found an associated risk of recurrent ear infections in children exposed prenatally to background levels of PCB and dioxins (51). More recently, it has been discovered that maternal PCB and TCDD dietary exposures reduced the measles vaccine response by age 3 and is associated with sex-dependent alterations in immune-specific transcripts in the cord blood (52, 53). Taken together, the developmental immunotoxicity of PCBs and dioxins suggests that through AHR-dependent, AHR-independent and sex-specific mechanisms, hematopoietic stem, and progenitor cells undergo reprogramming during development that ultimately impact the maturation and function of progeny cells required for maintaining a healthy immune system later in life.

Cigarette Smoking and Nicotine

Although smoking is widely recognized for its connection to an array of adverse health outcomes, its role as an endocrine disruptor is only recently emerging. For example, epidemiological studies suggest that the effects of smoking during pregnancy have a greater effect antagonizing male testosterone levels in the adult male offspring than occurs in men who only smoke as adults (54, 55). In addition to having probable endocrine-disrupting effects, smoking during pregnancy is developmentally immunotoxic demonstrated by a reduced antitumor immune response in pups developmentally exposed to the equivalent of a pack of cigarettes a day throughout gestation (56, 57). Specific to the hematopoietic system, adult smoking is associated with lower numbers of circulating CD34+ hematopoietic progenitor cells (58). Furthermore, a genome profiling study of cord blood and placenta hematopoietic progenitors from babies whose mother's smoked during pregnancy found a significant dysregulation of gene pathways involved in hematopoietic cell lineage differentiation, vascularization, inflammation, oxidative stress, and T-cell differentiation (59). Although not providing a direct functional assessment of the HSCs, the gene expression alterations found in children born to smokers are consistent with defects in hematopoiesis (59).

As with all smoking studies, a potential limitation is that the exposure to the fetus is that of a mixture of over 4000 compounds making mechanistic conclusions difficult. However, two studies of pregnant mice exposed exclusively to nicotine, at levels below the circulating concentrations measured in smokers, found significant changes to the hematopoietic system of the developing fetuses and neonates (60, 61). Specifically, there were decreases in self-renewal measured by in vitro Long-term initiating cell assays, a loss of HSC colonization of the fetal bone marrow, and reduced production of IL-6, a cytokine important for progenitor cell generation. Furthermore, expression of the nonneuronal nicotinic acetylcholine receptor was found in the fetal HSCs (60, 61). These data have public health implications not just for smoking during pregnancy but also on nicotine delivery systems used during pregnancy, including prescription patches, gums, and e-cigarettes.

Sex Steroid Receptor Regulation of Hematopoiesis

The effects of endocrine-disrupting compounds on hematopoiesis in general and fetal hematopoiesis specifically are very limited. However, recent findings on the important role played by sex steroid receptors on the regulation of hematopoiesis demonstrate this to be a potentially fertile area for future investigation. Specifically, estrogen has been demonstrated to adversely impact B- and T-cell lymphopoiesis (62) and androgens are known to promote erythropoiesis (63). Finally, estrogen receptor-α signaling promotes self-renewal, HSC proliferation, and erythropoiesis in female but not male mice (64). Taken together, these data raise the intriguing possibility that HSCs are sensitive targets of endocrine-disrupting compounds. What remains entirely unknown are the effects of hormone receptor signaling during fetal hematopoiesis, and thus, what are the potential effects of androgenic, antiandrogenic, estrogenic, or antiestrogenic compounds on development of the blood and immune system?

Pesticides

Identification of the hematopoietic system as a sensitive target of pesticide exposure comes from epidemiological studies and meta-analyses of farm workers that show an association between pesticide exposure and an increased risk for hematological malignancies (65, 66). These studies excluded pesticides potentially contaminated with dioxins and furans from their analyses; however, the specific pesticides that increase the risk could not be determined (65, 66). Although occupational exposures to pesticides are clearly of concern for these vulnerable populations and their families, an important consideration with potentially broader public health implications are the effects of daily low level exposure to pesticide mixtures through food intake. To address this question, a mixture of the pesticides alachlor, captan, diazinon, endosulfan, maneb, and mancozeb was administered to adult mice at the human acceptable daily intake levels for a period of 4 weeks. This exposure induced sex-specific alterations in HSC differentiation with a reduction in granulocyte/macrophage-colony forming unit assays in females and an increase in macrophage-colony forming unit assays in males (67). Human cord blood HSCs exposed to cypermethrin or mancozeb in vitro found that these pesticides to be erythro- and myelotoxic (68). Finally, exposure of mice to atrazine, chlorpyrifos, endosulfan, or the combination of all 3 throughout pregnancy, lactation and into adulthood found both sex-specific and pesticide-specific effects on colony forming unit assays (69). Notably, the mixture of all 3 pesticides antagonized the effects of individual pesticides on erythropoiesis and myelopoiesis, particularly endosulfan (69). Taken together, these data demonstrate that exposures to pesticides at levels at or below daily intake from food have impacts on hematopoiesis and these impacts are sex-dependent, consistent with the known endocrine-disrupting effects of some pesticides.

Future Challenges and Opportunities

It should be evident from this review that despite the knowledge that hormone receptors are expressed and functionally important for HSC self-renewal (64, 70), there is a significant gap between the hypothesis that blood cell progenitors are targets of endocrine-disrupting compounds and the existence of substantial data in either humans or rodent models in support of this hypothesis. Moreover, it is entirely unknown how hormone signaling and endocrine-disrupting compounds impact critical developmental events, including HSC emergence, migration, proliferation, energetic regulatory processes, and epigenetic coding that will permit self-renewal and multilineage differentiation throughout the duration of an individual's life. Thus, testing the impacts of environmental and endocrine-disrupting compounds during development of the fetal hematopoietic system represents a fertile area for future mechanistic research.

Among the important questions that must be considered with regards to endocrine-disrupting effects on hematopoiesis is determination of the susceptible generation to the environmental insult. Specifically, persistent low-level exposure to environmental toxicants during pregnancy may have multiple impacts on the mother, the fetus, or both. For example, an environmental exposure may antagonize the mother's normal endocrine signaling system impacting the levels, timing and duration of her production of factors necessary for maintaining the growth and maintenance of her developing baby. As an example of a maternal disorder impacting her baby, the pregnancy-induced insulin resistance associated with gestational diabetes results in higher levels of glucose transport across the placenta. The higher glucose raises the risk the baby will be born macrosomic. This excess fat storage increases risk for childhood asthma, obesity and type II diabetes as adults (71). Similarly, environmental exposures may impact the nutrient and oxygen exchange function of the placenta (72). Finally, many environmental contaminants are known to cross the placenta and as such directly expose and potentially impact the developing fetus. Identifying the mother, placenta, fetus, or combination as the target of endocrine-disrupting environmental exposures can be accomplished with carefully designed rodent studies using mice carrying gene targeted deletions or conditionally deleted genes specific for the receptors of a particular agent.

Identifying the principle target of individual environmental endocrine-disrupting compounds is crucial given the urgent need to expand our understanding of chemical mixtures on the developing fetus. Specifically, mixture experiments are needed because people are rarely, if ever, exposed to a single compound or even a single class of compounds. Thus, as we design and investigate the effects of mixtures, the experimental strategy should take into consideration 1) the most common mixtures found in the population of interest; 2) how the different constituents of the mixture impact the mother's pregnancy, her health and that of her developing baby; and 3) the potential for additive or antagonistic effects of the different mixture constituents on the developmental immune endpoint of interest. It is acknowledged that experimental studies involving mixtures add significant new layers of complexity, although scientists still do not fully understand all the mechanistic targets by which individual compounds contribute to DOHaD. Nevertheless, well-designed and executed studies may facilitate new partnerships between basic scientists and epidemiologists in order to facilitate experimental design that can advance this important field.

Perhaps the greatest scientific challenge for developmental immunotoxicologists is to be able to identify individuals and populations at risk for later-life immune diseases. Given the spectrum of diseases associated with an underlying immune dysfunction (73), improved treatment and prevention strategies are dependent on accurate biomarkers and functional assays of the immune progenitor cells. This challenge also presents an opportunity because although it may not be possible in the near term to identify a biomarker in a progenitor cell for a specific immune disease, given the sensitivity of progenitor cells to developmental insults, it may be possible to screen HSCs for their overall fitness at the time of birth and develop a predictive algorithm that portends potential risk for immunodeficiency. For example, elevated reactive oxygen species is known to adversely impact long-term self-renewal (13, 14) and as such could be measured in cord blood progenitors at birth as a diagnostic tool to potentially predict long-term consequences for hematopoiesis.

The benefits of this type of HSC fitness testing could yield tremendous societal benefits when the identification of biomarkers and risk factors for developmental disease susceptibility improves. Specifically, identification of populations vulnerable to later-life chronic health diseases will open up new approaches designed to prevent disease onset as opposed to the current focus on treatment and cures for already diagnosed disorders and diseases. Critically, prenatal dietary supplementation in mice has been shown to mitigate the effects of environmental stressors on epigenetic modifications, including those induced by the endocrine-disrupting compound bisphenol-A (74, 75). Prenatal dietary prevention approaches such as folate supplementation are already widespread in industrialized countries; however, the additional knowledge gained from biomarkers and cord blood analysis at birth has the potential to drive decision making and public health interventions that can be closely targeted to the most vulnerable populations.

In conclusion, as developmental immunotoxicologists continue to fill in the mechanistic data gaps on the effects of endocrine-disrupting and other environmental exposures on the developing immune system, there is an urgent need to improve predictive tools that can identify the developmental origins of immune health and disease. Using a multidisciplinary approach that combines traditional rodent models with rigorous exposure assessment and epidemiological population based studies has the potential to identify cellular and genetic biomarkers needed to predict and subsequently prevent an increased risk for later-life immune suppression. Leveraging these studies with functional assessment of hematopoietic progenitor cells present in the placenta and cord blood at the time of birth may advance our understanding of the environment's impact on human health and ultimately lead to improved prenatal care. By acting as the canary in the coal mine, fetal HSCs could be useful predictors of not just risk for later-life immune disease but also other maladies that have an underlying immune deficiency.