Obesity-driven disruption of haematopoiesis and the bone marrow niche

Abstract 

Obesity markedly increases susceptibility to a range of diseases and simultaneously undermines the viability and fate selection of haematopoietic stem cells (HSCs), and thus the kinetics of leukocyte production that is critical to innate and adaptive immunity. Considering that blood cell production and the differentiation of HSCs and their progeny is orchestrated, in part, by complex interacting signals emanating from the bone marrow microenvironment, it is not surprising that conditions that disturb bone marrow structure inevitably disrupt both the numbers and lineage-fates of these key blood cell progenitors. In addition to the increased adipose burden in visceral and subcutaneous compartments, obesity causes a marked increase in the size and number of adipocytes encroaching into the bone marrow space, almost certainly disturbing HSC interactions with neighbouring cells, which include osteoblasts, osteoclasts, mesenchymal cells and endothelial cells. As the global obesity pandemic grows, the short-term and long-term consequences of increased bone marrow adiposity on HSC lineage selection and immune function remain uncertain. This Review discusses the differentiation and function of haematopoietic cell populations, the principal physicochemical components of the bone marrow niche, and how this environment influences HSCs and haematopoiesis in general. The effect of adipocytes and adiposity on HSC and progenitor cell populations is also discussed, with the goal of understanding how obesity might compromise the core haematopoietic system.

Key Points

• Obesity represents a substantial public health challenge, including an increased burden of infectious diseases—suggesting the involvement of a compromised immune syste

• Haematopoietic stem cells reside in (and are regulated by) a complex, heterogeneous and tightly controlled microenvironment within the bone marrow

• The bone marrow environment undergoes considerable changes during obesity, including adipocyte hyperplasia and a phenotypic shift of adipocytes towards a white adipose profile

• Methods of limiting the progression of obesity and controlling its systemic and bone-marrow-based sequelae, such as exercise, might represent suitable approaches to maintain haematopoiesis and immune function



Introduction

 

There are now 1.7 billion overweight adults worldwide,¹ 300 million of whom are categorized as obese.² In the USA, data from the National Health and Nutrition Examination Survey estimated the overall prevalence of obesity in adults to be 35.7%.³ Alarmingly, obesity rates have increased most rapidly among young individuals,⁴ predisposing them to type 2 diabetes mellitus (T2DM),⁵ an increased lifetime risk of cardiovascular disease,⁶ cancer⁷ and early mortality,⁸ culminating in health-care costs of >$200 billion in the USA alone.⁹ Attempts to control either obesity¹⁰ or T2DM¹¹ by pharmacological intervention have been disappointing and plagued by off-target complications. In addition to the metabolic comorbidities traditionally associated with increased adiposity, obesity also increases an individual's susceptibility to infectious diseases¹² and catalyses a decrease in general health¹³—implying that obesity either directly or indirectly compromises immune function. Indeed, the progression of obesity, like that of other metabolic disorders such as anorexia, osteoporosis and ageing,¹⁴ is accompanied by an increase in bone marrow adiposity and disruption to the haematopoietic microenvironment, which presages damage to the immune system.
This Review seeks to explore the interactions between obesity, and haematopoiesis and the immune system through systemic effects and changes to the primary haematopoietic environment—the bone marrow niche. Firstly, the structure and function of the haematopoietic system is described by introducing the haematopoietic stem cell, its development and differentiation. Secondly, the heterogeneous regulatory environment of the bone marrow is discussed, with several key components described in detail. Thirdly, adipocytes and adipose tissue are described in states of both health and obesity. Fourthly, the available information on the effect of obesity on the bone marrow niche, haematopoiesis and immunity is presented. The article concludes with a focus on methods to suppress bone marrow adiposity.
 

Hematopoiesis

 

Haematopoiesis is the group of processes by which all formed elements of blood are created. Differentiated haematopoietic cells include erythrocytes, platelets and nearly a dozen types of leukocytes (white blood cells), which are responsible for both innate and adaptive immune surveillance and performance. Traditionally, haematopoietic differentiation is divided into either the myeloid lineage, which includes erythrocytes, granulocytes, macrophages and monocytes, and platelets, or into the lymphoid lineage, which includes B lymphocytes, T lymphocytes and natural killer (NK) cells. Although this model of haematopoiesis crudely identifies functional differences between mature cells, the initial descriptions were based on the location of cellular maturation (bone marrow versus lymph nodes and spleen), rather than on knowledge of the hierarchy of differentiation.¹⁵ With some acknowledged exceptions, myeloid cells are critical to innate immunity whereas lymphoid cells comprise the adaptive immune response.

Innate immunity is evolutionarily conserved in many organisms and enables a rapid defence against pathogens introduced through disease or injury.¹⁶ Although innate immunity can distinguish between host and foreign elements at the time an immune response is initiated, that response is not specific, and prolonged activation can readily lead to damage to healthy tissues.¹⁶ A suitable example is found in the experimental overexpression of one of the haematopoietic growth factors, granulocyte-macrophage colony-stimulating factor (GM-CSF), which leads to increased numbers and activation of macrophages as well as other leukocytes, resulting in massive inflammation and destruction of normal tissues.¹⁷ By contrast, adaptive immunity has the advantage of being highly specific in its targets, because of genetic rearrangements within cell-surface immune receptors and immunoglobulin proteins, and can reliably combat a much wider range of infectious agents.¹⁸ This specificity is achieved at the cost of rapid action, however, as lymphocytes of the adaptive immune system must first be primed to respond to specific antigens, at least when encountering a novel agent. Subsequent infections with the same antigen are met more quickly and forcefully with each exposure, as a consequence of immune memory.^{16, 18}

Haematopoietic Stem Cells

Haematopoietic stem cells (HSCs) are the developmental origin of the haematopoietic system. These cells are primitive, pluripotent progenitors that differentiate, in part, in response to environmental signals that mediate homeostatic maintenance of the entire haematopoietic system, as well as when increased blood cell production is required in times of enhanced demand, such as in response to infection or enhanced blood cell destruction.^{19, 20} The balanced and timely regulation of HSC differentiation is critical for maintaining normal immune function. The notion that HSCs are the developmental source of the haematopoietic system originated in the early part of the 20^th century,²¹ but was not proven until the 1960s by lethal irradiation and bone marrow transplantation studies.²² To this day, HSCs are still functionally defined as being capable of rescuing a lethally irradiated animal and restoring stable, long-term and full-spectrum haematopoiesis.^{23, 24}

The hierarchy of haematopoietic differentiation, with regard to the functional and phenotypic identity of sequentially generated populations, and how differentiation 'fate' is achieved, is an active area of research. The transitions at the origin of this hierarchy are self-evidently important to the haematopoietic system as a whole. The existence of phenotypically distinct subpopulations that maintain haematopoietic reconstitution for differing durations within the progenitor population capable of producing the full spectrum of haematopoietic progeny cells has long been recognized.²⁵

 

The presence of long-term HSCs (LT-HSCs) and short-term HSCs (ST-HSCs) in human bone marrow is of clinical relevance when considering the efficacy of therapeutic bone marrow transplantation.²⁶ In most models of haematopoiesis, HSC multipotency is coupled with the ability to self-renew.²⁷ In such models, ST-HSCs are considered an intermediary state in which some 'stem-like' properties are retained until the cell commits to one of the principal (myeloid or lymphoid) lineages.²⁸ These lineage decisions, which are influenced by cytokine signalling,²⁹ metabolism³⁰ and stochastic processes,³¹ occur during maturation, after the loss of 'stemness'. This concept has changed lately, with the recognition that even primitive HSCs can be biased in their differentiation profiles.³² Interestingly, the clonal bias of certain HSC subpopulations has been identified in both ST-HSCs and LT-HSCs.²⁷ In this model, differentiation of LT-HSCs generates ST-HSCs and myeloid progenitors, whereas lymphoid (and continued myeloid) potential derives from ST-HSCs.²⁷ In this hypothesis, there are no distinct HSC populations dedicated to maintaining a certain lineage; rather, the lineage potential changes as HSCs mature.

 

Adult HSCs arise from the dorsal aortic section of the aorta–gonad–mesonephros region at embryonic day 10.5 in the mouse.^{24, 33} These cells then migrate to populate the fetal liver, from which they later migrate and home to the spleen and most importantly the bone marrow,³⁴ where adult HSCs reside and their progeny develop. Ultimately, the relationship between the bone marrow and the haematopoietic populations is dependent on the composition of the bone marrow. Consequently, changes to the bone marrow phenotype, such as those that occur with increased adiposity, inevitably affect haematopoietic maintenance and differentiation, ultimately putting immune function at risk.

Paracrine and endocrine regulation

A number of haematopoietic cytokines have been identified that contribute to regulation, survival, proliferation and differentiation of the bone marrow. Whilst this Review focuses on the bone marrow microenvironment and cell–cell interactions, a brief overview of the effects of soluble growth factors on HSCs is provided.

 

Although humoral regulation of erythropoiesis was first described in the early 1900s,³⁵ it was not until the 1950s that erythropoietin was confirmed as the primary regulator of red cell production.³⁶ In the 1960s and 1970s, the development of in vitro assays to measure growth factors secreted by bone marrow cells that produced leukocytes identified additional humoral regulators of leukocyte and megakaryocyte (platelet) production.³⁷ Even though >30 such factors have now been identified, only a few cytokines regulate HSCs and only two of these have essential and non-redundant roles: Kit ligand (also known as stem cell factor or SCF)³⁸ and thrombopoietin.³⁹ Both of these cytokines, which are produced in peripheral organs and by bone marrow stromal cells, act by binding to their cognate receptors on the cell surface of HSCs and stimulating intracellular kinase activity that leads to the activation of several signalling pathways involved in cell survival and differentiation. Not surprisingly, given these properties, the receptors for both SCF⁴⁰ and thrombopoietin⁴¹ are involved in neoplastic bone marrow diseases.

The bone marrow niche

In adult individuals, the bone marrow is the primary site of haematopoiesis. The bone marrow microenvironment, or niche, provides a critical regulatory milieu that is responsible, in part (along with the stochastic increased and decreased expression of lineage-specific transcription factors and haematopoietic cytokines generated both locally and from afar), for orchestrating the differentiation of HSCs into the different types of mature blood cells.⁴² Among all niche residents, HSCs are actually quite rare, comprising ~0.001% of total bone marrow cells.⁴³ The remaining cells are haematopoietic progenitors at various stages of maturity, nearly mature blood cells and nonhaematopoietic cells that provide key regulatory signals in the haematopoietic niche. The fact that heterogeneous populations develop in the niche denotes that changes in the composition of the niche can affect haematopoiesis at multiple developmental levels and via a variety of mechanisms. A brief overview of the inherent complexity of the niche, both in terms of its anatomy and signalling pathways, structured around several of the regulatory influences of resident cell types and their capacity to bias the activity and fate of different haematopoietic populations follows.

Complexity

Osteoblasts

The endosteal bone surface has long been recognized as the principal component of the haematopoietic niche, with osteoblasts, osteoclasts and mesenchymal stem cells (MSCs) all influencing HSC fate selection (Figure 1).⁴² Although the primary function of osteoblasts is to secrete nonmineralized bone matrix (osteoid) in the phase preceding bone mineralization, osteoblasts located on the endosteal surface also have a major role in the regulation of HSCs. Osteoblast lineage cells immobilize HSCs by, among other mechanisms, the binding of integrin α₄β₁ to VCAM1, which reduces apoptosis and induces HSC quiescence.⁴⁴ Similarly, soluble stromal-cell-derived factor 1 (also known as CXCL12) and angiopoietin-1 secreted by osteoblasts prevents HSC mobilization from the niche and promotes cell quiescence.^{45, 46} Osteoblasts are also closely coupled to HSC proliferation, as increases in the osteoblast population lead to concomitant increases in HSC numbers.^{47, 48} This expansion of the HSC compartment seems to be mediated by osteoblastic Notch signalling,⁴⁸ as osteoblastic expression of Notch ligands⁴⁹ leads to proliferation of both ST-HSCs and LT-HSCs. However this role might be limited to haematopoietic recovery and not homeostasis per se.⁵⁰ Other factors involved in the osteoblast lineage regulation of HSC numbers include osteopontin,⁵¹ Wnt, N-cadherin, thrombopoietin⁵² and angiopoietin.⁵³ Nevertheless, a change in the number or activity of osteoblasts, which invariably occurs with ageing,⁵⁴ diabetes⁵¹ or obesity,⁵⁵ will ultimately have an effect on HSCs. Interestingly, all these conditions are characterized by increased inflammation, which harms both osteoblast proliferation and function,^{54, 56} as well as the maintenance of HSCs^{57, 58}—further highlighting the delicate interaction between these cell populations

 

Figure 2: Haematopoietic stem cell differentiation.

HSC differentiation in bone marrow of a healthy (top right) and obese mouse (bottom right), as represented at the proximal tibial metaphysis. HSCs differentiate to produce all blood lineages under niche-based regulatory control. In homeostasis, differentiation is balanced in order to maintain immune function and to replenish erythrocytes. In states of obesity, the adipocyte phenotype shifts from a hybrid brown and white (beige) profile towards a classical white adipose profile. The bone marrow becomes entangled with these adipocytes, which produce excess concentrations of adipokines and cause inflammation. This inflammation results in a proinflammatory bias and a loss of leukocyte balance, in which some myeloid lineage populations are increased, seemingly at the expense of lymphoid populations. This loss of lymphopoiesis and the proinflammatory state of the immune system are likely to contribute to the immune deficits observed in states of obesity. Abbreviations: CLP, common lymphoid progenitor; CMP, common myeloid progenitor; HSC, haematopoietic stem cell; LT, long term; MPP, multipotent progenitor; NK, natural killer; ST, short term.

 

Ultimately, any conclusions regarding the role of adipocytes in the bone marrow niche must be approached cautiously, owing to issues in extrapolating data from studies of mice to humans. This statement is particularly true considering the different patterns of fat accretion between the two species. Specifically, whereas ageing accelerates bone marrow adiposity in both mice and humans, sites of red bone marrow (for example in the femur) become highly adipocytic long before senescence in humans, whereas comparable mouse bone marrow remains relatively adipocyte-free until advanced old age.^{149, 150} These different rates of adipocyte accumulation indicate the potential for differences in the physiological roles and activity of adipocytes between humans and mice. At the very least, these observations emphasize the conclusion reached earlier, which regards the presence of adipocytes as not necessarily disruptive, particularly in the absence of a phenotypic shift to bone marrow adipocytes (as has been observed in advanced ageing and metabolic disorders such as obesity and diabetes mellitus).^{106, 114}

Suppression of bone marrow adiposity

The potential negative effect of increased adipogenesis on bone marrow organization and haematopoietic function highlights a critical need for strategies that safely and effectively limit or reverse adipocyte invasion into the bone marrow space. Although reduced caloric intake remains a key mechanism for controlling adiposity in patients with obesity, exercise also fosters weight loss^{151, 152} and improves physical performance and physiological function.¹⁵¹ Indeed, exercise can directly address many of the health consequences of obesity, without first reducing adiposity. For example, exercise improves cardiac function in a dose-dependent manner,¹⁵³ reduces blood pressure in individuals with hypertension¹⁵⁴ and can improve serum markers of liver function and suppress hepatocyte apoptosis in patients with nonalcoholic fatty liver disease, before altering body composition.¹⁵⁵ Bone marrow adiposity, assessed as bone marrow density using MRI in female athletes, can also be proportionally reduced by high-impact loading regimes.^{156, 157} Similarly, in mice fed a high-fat diet, bone marrow adiposity (shown by osmium tetroxide staining) was greatly increased compared to that in lean control mice. However, voluntary use of a running wheel significantly decreased the level of bone marrow adiposity in both lean animals and those fed a high-fat diet.¹⁵⁸ Interestingly, protection of the bone marrow phenotype in obese animals was achieved without reducing systemic adiposity,¹⁵⁸ indicating that this process might be related more to the influence of mechanical stimulation on bone marrow adipocytes and the niche as a whole, rather than to changes in energy balance. Such a relationship is highlighted by the association of reduced bone marrow adiposity with concomitant increases in skeletal parameters in these studies.^{156, 158}

 

The mechanisms responsible for the reduction of bone marrow adiposity are complex but, considering the inverse relationship between bone marrow adiposity and skeletal mass, might involve a competitive balance between osteoblast and adipocyte differentiation. One key factor in this process is growth hormone. During ageing, the bone marrow becomes gradually filled with adipocytes and bone mass is lost, and levels of growth hormone also decrease. Impairment of growth hormone production in mice led to an accumulation of adipocytes in the bone marrow despite a reduction in body weight;¹⁵⁹ the increased levels of bone marrow adipocytes were normalized and osteoblast activity was increased by growth hormone replacement.¹⁵⁹ Importantly, growth hormone is produced in response to exercise,^{160, 161} providing some insight into how exercise influences the bone-marrow microenvironment. Similarly, whereas PPAR-γ is required for adipocyte differentiation, its absence leads to elevated osteoblastogenesis.¹⁶² Treadmill running in rats prevents ovariectomy-induced bone loss by limiting PPAR-γ expression, whilst simultaneously suppressing adipocyte encroachment into the bone marrow.¹⁶³ Supporting this interaction, MSCs subjected to very subtle mechanical signals in vivo are biased away from adipogenesis and towards osteoblastogenesis,¹⁶⁴ even in highly adipogenic environments.¹⁶⁵ Likewise, in vitro stretching of MSCs reduced PPAR-γ signalling and adipogenesis, even during PPAR-γ activation.¹⁶⁶

 

Another important and mechanosensitive regulatory mechanism in the stem cell niche is β-catenin signalling. Diminished levels of β-catenin are a hallmark of adipogenesis in MSCs, whereas mechanical stretching increases levels of β-catenin and inhibits adipogenesis in vitro.^{167, 168} Functional β-catenin signalling and its activation of the canonical Wnt pathway are also necessary for haematopoietic homing to the bone marrow niche¹⁶⁹ and engraftment of ex vivo expanded HSCs.¹⁷⁰

 

Furthermore, β-catenin is necessary for haematopoietic regeneration following injury.¹⁷¹ The level of β-catenin signalling within HSCs themselves is dependent on niche health, as HSCs isolated from diabetic mice have reduced β-catenin expression, which can be rescued through co-culture with stromal cells from nondiabetic mice.⁵¹ The importance of β-catenin signalling to haematopoiesis, its inverse relationship to adiposity and its sensitivity to mechanical interventions, suggest that the canonical Wnt pathway is a link between exercise and haematopoietic health. Cumulatively, these studies demonstrate the potential of exercise, or its surrogates, to bias MSC fate selection and thus restrict bone marrow adiposity and simultaneously stimulate osteoblastic activity, ultimately leading to the promotion of haematopoietic function.¹²⁴

Conclusions

The bone marrow niche regulates haematopoiesis through complex and sometimes compensating influences. The targets of this regulation include not only the most primitive HSCs, but also leukocytes transitioning through a range of developmental stages. These influences are delivered via cell–cell contact, soluble signalling factors and even changes in the local physicochemical and inflammatory environments. The delicate interactions of cellular, physical and chemical factors are, in aggregate, responsible for balancing haematopoietic quiescence, proliferation and differentiation. As such, changes to the niche composition can have profound effects on haematopoiesis and the proper functioning and robustness of the immune system.

 

Obesity is thought to increase bone marrow adiposity (in some cases rapidly)^{110, 111, 112, 139, 158}—a condition that can disrupt both chemical and physical connections, as well as the cellular and signalling content of this fragile but vital niche. Such disruption results in diminished lymphocyte populations,^{123, 124, 139} immune function,^{125, 126} and ultimately the suppression of HSC proliferation.^{114, 143} In addition to its effect on the bone marrow, obesity also increases susceptibility to a range of chronic comorbidities, including increased inflammation related to the increase in systemic adiposity. As yet, the relative consequences of obesity for the haematopoietic system, as caused by systemic changes versus local shifts in the niche, cannot be defined. Nevertheless, considering other conditions of elevated bone marrow adiposity, such as ageing and anorexia,^{14, 84} we can safely conclude that the increased adiposity observed in individuals with obesity can quickly lead to profound haematopoietic insult and a compromised immune system. In conclusion, these results, ranging from signal transduction studies to bone marrow phenotypes, emphasize the importance of the bone marrow composition on overall health of the immune system and the organism itself, and advocate strategies such as exercise to limit the consequences of excess bone marrow adiposity.