Immune dysfunction in developmental programming of type 2 diabetes mellitus

Abstract

Intrauterine growth restriction (IUGR) is a common complication of pregnancy and increases the risk of the offspring developing type 2 diabetes mellitus (T2DM) later in life. Alterations in the immune system are implicated in the pathogenesis of IUGR-induced T2DM. The development of the fetal immune system is a delicate balance as it must remain tolerant of maternal antigens whilst also preparing for the post-birth environment. In addition, the fetal immune system is susceptible to an altered intrauterine milieu caused by maternal and placental inflammatory mediators or secondary to nutrient and oxygen deprivation. Pancreatic-resident macrophages populate the pancreas during fetal development, and their phenotype is dynamic through the neonatal period. Furthermore, macrophages in the islets are instrumental in islet development as they influence β-cell proliferation and islet neogenesis. In addition, cytokines, derived from β-cells and macrophages, are important to islet homeostasis in the fetus and adult and, when perturbed, can cause islet dysfunction. Several activated immune pathways have been identified in the islets of people who experienced IUGR, with alternations in the levels of IL-1β and IL-4 as well as changes in TGFβ signalling. Leptin levels are also altered. Immunomodulation has shown therapeutic benefit in T2DM and might be particularly useful in IUGR-induced T2DM.

Over the past three decades, it has been increasingly recognized that the risk of adult health disorders, particularly type 2 diabetes mellitus (T2DM), can be influenced by prenatal and infant environmental exposures (that is, developmental programming). Intrauterine growth restriction (IUGR) is associated with a considerable risk of T2DM, obesity and cardiovascular disease in the offspring. Interestingly, the animal models of IUGR all have a similar β-cell phenotype, which suggests that, similar to the effects of IUGR in humans, impaired β-cell function has a central role in the development of T2DM^1,2. As will be discussed in this Review, reports suggest that immune pathways are altered in offspring who experienced IUGR, which might have a role in the pathogenesis of T2DM. In this Review, we will discuss the susceptibility of the fetal immune system to IUGR and the consequences on pancreatic islet development. We will also highlight that cytokines are altered by IUGR and their known involvement in T2DM. Finally, we will discuss the potential for immunomodulation as a therapeutic approach to treat IUGR-induced T2DM.

Small for gestational age (SGA).

Birth weight that is below the 10th percentile.

Evidence for fetal origin of T2DM

The fetal origin of disease hypothesis, first posed by Professor David Barker, postulates that exposure to an adverse intrauterine environment has the potential to influence adult health and disease^3–5. The period from conception to birth is a time of rapid growth, cellular replication and differentiation, and functional maturation of organ systems. These processes are very sensitive to alterations in nutrient availability, and an abnormal intrauterine metabolic milieu can have long-lasting effects on the offspring. Perhaps the best example of how nutrient availability during pregnancy affects long-term health and disease in the offspring is the Dutch famine of 1944–1945. A landmark cohort study evaluated 300,000 men born during the Dutch famine and found that offspring whose mothers experienced famine conditions during the first half of their pregnancy had an increased incidence of obesity at 19 years old⁶. Subsequent studies correlated low birth weight with an increased incidence of cardiovascular disease³ as well as increased systolic blood pressure and triglyceride levels⁷. The incidence of T2DM was also increased in those born small for gestational age (SGA)^8–10. In fact, men who were born 2.5 kg or smaller had seven times more risk of impaired glucose tolerance or T2DM compared with those who had a birth weight average of 3.5 kg (REFS^8–10). Since the inception of the fetal origin of disease hypothesis, many studies have replicated the relationship between low birth weight and adult disease^11–20.

Lifestyle factors also have a notable effect on the development of disease. Through many studies that will not be discussed in this Review, we know that parental socioeconomic status, smoking status, physical activity levels, diet and occupation influence offspring health^21–24. In the past few years, epidemiologic studies have controlled for these confounding variables. For instance, the Nurse’s Health Study found that women born with low birth weight had an increased risk of coronary heart disease, stroke and T2DM even when lifestyle risk factors were accounted for²⁵. Another study evaluated 22,000 men from the USA and found that those with a birth weight of 5.5 pounds or less had an increased incidence of hypertension and T2DM in adulthood compared with men with an average birth weight, independent of lifestyle factors¹¹.

Fetal development is rapid and the gestational timing of the insult determines the effects on development. Epidemiological studies have correlated gestational timing with specific outcomes. For instance, if growth is restricted during the first trimester in fetal development, offspring are at a higher risk of coronary heart disease and atherogenic lipid profiles compared with individuals who were born average for gestational age^26,27. However, growth restriction in the second and third trimesters in gestation results in impaired glucose tolerance and T2DM in adulthood²⁸. These different windows of susceptibility might point to the mechanism by which fetal growth restriction leads to adult disease.

Uteroplacental insufficiency is now the most common cause of impaired fetal growth²⁹. Conditions such as preeclampsia, chronic hypertension and smoking result in uteroplacental insufficiency and are linked to the subsequent development of hypertension, T2DM and obesity in the offspring^30–32. Of importance, these diverse maternal disorders are all associated with dysregulation of the fetal and neonatal immune system^33–39.

The fetal immune system

Fetal immune development.

The development of the fetal immune system is a delicate balance that is susceptible to alterations of the intrauterine milieu. In utero, the fetus is exposed to maternal proteins that have the potential to trigger an immune reaction^40–43 (FIG. 1). Tolerance to maternal antigens is therefore required and is achieved by fetal immune tolerance⁴⁴. Regulatory T cells are the dominant cell type that promotes tolerance. A high proportion of leukocytes in the fetus are regulatory T cells but the number of these cells dramatically decreases after birth^45,46. Other immune cell populations also alter phenotype and function after birth. For instance, T cells switch from a γδ T cell receptor to an αβ T cell receptor after birth⁴⁷. B cells also shift from a more primitive B1 state to the B2 state after birth⁴⁸. Importantly, T cells and macrophages promote tolerance in the fetus and acquire a pro-inflammatory phenotype after birth^49,50. The dramatic changes to the immune system at birth probably reflect the change in environment and potential pathogen exposure.

Fig. 1 |. Fetal immune development.

During a normal pregnancy, the fetus is exposed to maternal antigens, with the potential to elicit an immune response. To prevent inappropriate inflammation, the fetal immune system promotes tolerance of foreign antigens. At birth, the neonate is exposed to the environment and potential pathogens. In the neonate, the immune system shifts towards a pro-inflammatory phenotype. During these developmental windows, the phenotypes are determined by the dominant immune cell populations.

Uteroplacental insufficiency.

A complication of pregnancy when the placenta is unable to deliver an adequate supply of nutrients and oxygen to the fetus.

The development of the immune system starts as early as week 3 of gestation in humans or embryonic day 7 in mice. Derived from the ectoderm, the first immune cells populate in the yolk sac^51,52. At 5 weeks of gestation, the fetal liver is formed and haematopoietic stem cells give rise to myeloid, lymphoid and erythroid cells^51,52. The fetal liver remains the primary site of myelopoiesis, until 20–24 weeks of gestation in humans and at birth in mice, at which point the bone marrow is colonized⁵³. The bone marrow remains the primary site of haematopoiesis through adult life. Interestingly, tissue-resident macrophages populate specific tissues during fetal development^54,55. Depending on the tissue, resident macrophages can arise from the yolk sac, fetal liver or bone marrow. The pancreatic-resident macrophage populations will be discussed in more depth in subsequent sections.

The in utero milieu affects neonatal immunity

The immune system is altered in offspring affected by IUGR, both in humans and animal models^56–62. Systemic and organ-specific alterations in immune cells have been reported and are associated with organ dysfunction^57–60,63. While these data are limited, they suggest a potential underlying mechanism of pathology attributed to IUGR.

Clinically, growth restriction during fetal development leads to altered neonatal immunity. Neonates born SGA have alterations of their immune system, as demonstrated by the increased incidence of post-natal infections in these children⁶⁴. While a portion of the elevated risk can be attributed to the increased need for post-natal care in a hospital setting, the immune system of neonates born SGA is also altered. For instance, neonates born SGA have fewer circulating leukocytes^59–61 and isolated natural killer and Vγ9Vδ2 T cells have reduced antiviral capacity compared with those born average for gestational age^59,60. Additionally, the thymus is smaller in neonates born SGA than in neonates of normal birth weight^65,66. While these data suggest that the pro-inflammatory response is dampened in infants born SGA, which results in increased susceptibility to infection, the systemic immune alterations have yet to be well characterized or the mechanism of alteration identified.

In animal models of IUGR, alterations in the immune system have been alluded to but not extensively studied. In a sheep model of IUGR, the expression of genes with protein products involved in the immune system, namely those encoding components of TGFβ signalling, was altered in fetal islets of sheep exposed to IUGR compared with those of control sheep⁶³. Additionally, in a pig model of IUGR, levels of IL-10 were increased in the ileum of the offspring, which was associated with increased intestinal permeability and dysfunction⁵⁷. Similar to reports in humans, circulating lymphocyte proliferation and cytokine production were also systemically reduced in piglets that had experienced IUGR^56,62. IUGR in rats is also associated with alterations in fetal and neonatal immune function. For instance, serum levels of the chemokines MCP1 and RANTES are increased at postnatal day 14 in rats exposed to IUGR compared with control rats⁵⁸. Additionally, levels of the chemokines and cytokines IL-4, IL-2, IL-10, RANTES, eotaxin and MCP1 are elevated in fetal islets of offspring exposed to IUGR, with some attenuation by 2 weeks of age. While these alterations in the immune system induced by IUGR have been reported, the mechanisms have yet to be explored in these models.

Parabiosis studies.

A laboratory technique to study physiology whereby two living organisms are joined together surgically to develop a single, shared physiological system.

Cumulatively, these studies demonstrate that the fetal immune system is susceptible to an altered intrauterine milieu. In the past few years, the role of inflammation in the pathogenesis of metabolic diseases has been of great interest, primarily from the viewpoint of immunometabolism. Thus, it is critical to elucidate the role of inflammation during fetal and early neonatal life in the pathogenesis of IUGR-induced T2DM in adulthood.

Pancreatic-resident macrophages

Macrophage derivation and phenotype.

Specific macrophage populations are resident in the exocrine and endocrine portions of the pancreas. Their location in the pancreas determines their origin and turnover rate. Macrophages are first observed in the mouse pancreas at embryonic day 14.5 (REF.⁶⁷). As this time point is before blood vessel formation, macrophage precursors are probably already present in the pancreas. In fact, F4/80+ cells have been collected from excised pancreata from embryonic day 12.5 mouse embryos after the sample was cultured for 5 days with macrophage colony-stimulating factor⁶⁷. Macrophages continue to populate the developing pancreas and are present in islets and the exocrine compartment in mice by 1 week of age⁶⁸. Early population of the pancreas with macrophages suggests that resident macrophages are susceptible to fetal perturbations.

Lineage tracing in mice suggests that adult islet macrophages are derived from adult haematopoietic stem cells. There are two exocrine macrophage populations in the adult, which are derived from the yolk sac or adult haematopoietic stem cells⁶⁸. Despite their derivation from adult haematopoietic stem cells, islet and exocrine macrophages are long lived, as demonstrated by parabiosis studies⁶⁸. Following injury, such as irradiation, monocytes from the blood replace macrophage populations in the pancreas⁶⁸. However, monocyte-derived replacement is not dependent on C-C chemokine receptor 2 (CCR2) as CCR2^−/− mice have a similar maintenance of islet and stromal pancreatic macrophages as wild-type mice⁶⁸. The mechanism by which macrophages traffic to the pancreas at steady state has yet to be fully elucidated⁶⁸. Additionally, some in situ proliferation occurs following injury, which suggests that macrophage precursors might be present in the pancreas⁶⁸.

The microenvironment of pancreatic-resident macrophages determines their phenotype. Leukocyte populations in the mouse exocrine and endocrine pancreas have been thoroughly characterized⁶⁸. In the exocrine pancreas, several leukocyte populations are dominated by F4/80⁺CD11b⁺CD11c⁺MHC-II⁺ macrophages⁶⁸. This Review will focus primarily on immune cells resident to the endocrine pancreas.

Macrophages are resident to pancreatic islets (FIG. 2). Early reports in mice suggested that this myeloid population was comprised of dendritic cells owing to their high expression of CD11c and MHC-II⁶⁹. As the definition of a macrophage has been expanded and the characterization of islet macrophages further elucidated, these CD11c⁺MHC-II⁺ cells are now referred to as macrophages^68,70. Islet macrophages express F4/80, CD11c, MHC-II, CD64, CX3CR1 and lysozyme⁶⁸. Unlike macrophages found in the exocrine pancreas, they do not express CD206 or galactose-type C-type lectin⁶⁸. The transcriptome of flow sorted mouse islet macrophages demonstrates a skewing towards the pro-inflammatory M1 phenotype, as shown by the high expression levels of TNF and IL-1β⁶⁸. In the exocrine pancreas, macrophages are more skewed towards the M2 phenotype and express ARG1, IL-10, FIZZ1 and YM1 (REF.⁶⁸). Islet macrophages also express genes encoding co-stimulatory molecules, such as CD86, PVR, galectin 9 and PDL1, genes encoding the components of toll-like receptor and myeloid cell chemokines⁷¹. Uniquely, these macrophages express genes that indicate that they have lysosomal function⁷¹. The expression of such genes and proteins suggests that these macrophages are primed to sense their environment and to potentially act as antigen-presenting cells.

Fig. 2 |. Pancreatic immune development.

Macrophages populate the pancreas during embryogenesis, and their phenotype continues to develop during the perinatal period. Macrophages are influenced by their environment and adopt different phenotypes in the endocrine (that is, islets) compared with the exocrine tissue.

Phenotype transitions in the neonate.

The neonatal period is dynamic and the resident macrophage population shifts from the fetal to the adult phenotype. MHC-II expression is low at 1 week of age and increases to adult levels by 4 weeks of age in mouse islet and exocrine pancreas macrophages⁶⁸. Our previous study in rats also supports an adaptation of immune cell populations after birth. Gene transcripts measured by microarray highlight elevated activity of type 2 T helper cell pathways at embryonic day 19 that decrease by postnatal day 14 in normal rat islet lysates⁵⁸. This finding suggests that the macrophage populations, though seeded during fetal development, do not acquire their adult phenotype until the neonatal period.

The role of macrophages in islet development.

Several studies have now shown that macrophages have a major role in pancreas development. Early research using the Csf1-knockout (Csf1^op/op) mouse demonstrated that macrophages are important for islet formation. Banaei-Bouchareb et al.⁷² first identified an islet phenotype in the Csf1^op/op mouse. They found that, at embryonic day 18.5 and postnatal day 21, Csf1^op/op mice had smaller islets but there was no difference in β-cell mass compared with wild-type mice. However, while islets were smaller and fewer large islets were found in Csf1^op/op mice compared with wild-type mice, the overall number of islets was increased, which suggests an arrest in β-cell proliferation⁷². Interestingly, pancreatic duct cell proliferation was increased in the Csf1^op/op mouse at postnatal day 21, which suggests increased islet neogenesis⁷². Alterations in the islet size and number have since been confirmed in the Csf1^op/op mouse^68,69.

As expected, the number of macrophages in the Csf1^op/op mouse is reduced. In these mice, 85% of islets had no macrophages and the remaining had significantly fewer macrophages than those of wild-type animals^68,69. In islets, the number of macrophages was reduced overall, but the phenotype of those that remained was also altered. Monocyte precursors (F4/80⁻CD11b⁻) were also present in Csf1^op/op mouse islets⁶⁸. A dramatic loss of the CD206⁺ macrophages occurs in the exocrine pancreas, but CD206⁻ macrophages remain the same in the Csf1^op/op compared with wild-type animals⁶⁸. The authors did not report a phenotype in the Csf1^op/op mice. This loss of particular macrophage populations implicates a major role for these macrophages in islet formation. To support the direct role of macrophages in islet development, the effect of macrophage colony-stimulating factor on pancreas explants taken from fetal mice (embryonic day 12.5) has been tested⁶⁷. Macrophage colony-stimulating factor increased the number of insulinpositive cells fourfold and glucagon-positive cells trended towards an increase as well. Together, this research suggests that macrophages have a key role in β-cell development and function (FIG. 3).

Fig. 3 |. Macrophages influence islet development.

Macrophages are instrumental to the maturation of pancreatic islets. a | In the adult, macrophages populate the islets and exocrine pancreas. b | In mice lacking macrophage colony-stimulating factor, there is a reduction of macrophages in the islets and of CD206⁺ macrophages in the exocrine pancreas. Islets of these mice are also smaller in size than those of wild-type mice, though more numerous.

Macrophages sense the environment

Extensive research in adult mouse islets has identified a crucial relationship with macrophages that is responsible for islet health and function. Islet macrophages are influenced directly by their microenvironment and indirectly via systemic exposure (FIG. 4). β-Cells release granules, proteins and cytokines that have direct effects on islet macrophages. Islet macrophages also respond to changes in the extracellular matrix. Finally, systemic signals are known to influence islet health and function. Through an understanding of this relationship in adults, we can potentially identify susceptible pathways important to islet function and development in the neonatal period.

Fig. 4 |. Resident macrophages sense their micro and systemic environments.

Islet-resident macrophages respond to intra-islet and systemic signals such as adipokines and cytokines (inflammatory signals). They in turn release cytokines (such as TGFβ and IL-1β) that influence β-cell function. These dynamic interactions are crucial to islet function and have the potential to cause pathology.

β-Cell communication with macrophages.

β-Cells are known to directly communicate with islet macrophages in mice. Microscopy has been used to demonstrate that islet macrophages are in close proximity to β-cells and take up the granules directly from β-cells^73,74. Intact granules are different from those released via insulin-stimulated glucose release, suggesting an intimate connection between macrophages and β-cells⁷³. Dense core granules from β-cells have been found in islet macrophages and are presented to T cells⁷⁴. This connection could serve as a method of direct communication between β-cells and macrophages.

β-Cells can also produce and release chemicals into the microenvironment that lead to immune cell activation. The promoter region of the gene that encodes islet amyloid polypeptide (IAPP; also known as amylin) shares similar control elements to the gene that encodes insulin^75,76, which suggests synchrony in their expression. IAPP has been found in macrophages residing in the islets of humans with T2DM and non-human primates with diabetes mellitus⁷⁷. Interestingly, this finding can precede the development of hyperglycaemia and insulin resistance in non-human primates⁷⁸. Both a high-fat diet and hyperglycaemia increase levels of IAPP in mice^79,80. In mice fed a high-fat diet, this increase in IAPP levels leads to increased levels of IL-1β in the islet⁸⁰ and to activation of the inflammasome^79,80. When macrophages are depleted, levels of IL-1β return to normal and glucose-stimulated insulin secretion is improved. The accumulation of IAPP in the islets elicits an immune response that can lead to T2DM^79–81.

Macrophages sense circulating factors.

In mice, macrophages are often found near blood vessels⁶⁹ and sense circulating factors via the extension of filopodia into the lumen of blood vessels⁸². Islet macrophages can engulf microparticles in the blood⁷³ and are activated in response to lipopolysaccharide in the blood⁷¹. Owing to the close relationship between islet macrophages and the systemic circulation, islet macrophages might be a signal for T cell infiltration in type 1 diabetes mellitus (T1DM) and in fact express the β-cell–peptide–MHC complex⁸³. The response to systemic inflammation is key to the development of T1DM and might also be important in the development of T2DM.

Immune cells in the pancreas can also sense altered systemic immunity. Obesity elicits a chronic inflammatory state and signals from adipose tissue enter the circulation, which results in pancreatic inflammation. Indeed, macrophages have been reported to drive islet inflammation in obese mice⁸⁴. Interestingly, macrophages proliferated locally rather than being recruited from the circulation⁸⁴, highlighting the immediate local response to systemic signals. The role of obesity-driven macrophages in islet inflammation and β-cell function has been reviewed elsewhere⁸⁵.

Islet cytokines

β-Cells also release cytokines that directly activate macrophages. Cytokines influence islet development, homeostasis and subsequent glucose control. Two crucial cytokines necessary for islet health are IL-1β and TGFβ. Interestingly, both have also been implicated in IUGR.

IL-1β influences islet function.

Levels of IL-1β, a key proinflammatory cytokine, are increased in islets of people with T2DM and a great deal of research has been done to understand its role in islet function. Several cell types in human islets express IL-1β^86–88. Macrophage-derived IL-1β has been found in rodent models of T1DM, and, in fact, IL-1β is responsible for the inhibition of glucose-stimulated insulin secretion in response to inflammatory stimulators such as TNF and lipopolysaccharide⁸⁷. Pancreatic α-cells also serve as a source of IL-1β but levels of α-cell-derived IL-1β are similar between islets from humans with and without T2DM⁸⁶. However, β-cells from patients with T2DM have a higher expression of IL-1β mRNA than β-cells from healthy controls^88,89.

Owing to its increased expression in islets of patients with T2DM, researchers have studied the effects of IL-1β on islet function. IL-1β expression is increased in response to elevated levels of glucose⁸⁸ or saturated fatty acids⁹⁰ in humans. Depending on the level of IL-1β, the effect on β-cells varies. At low concentrations or short-term increased levels of IL-1β, glucose-stimulated insulin secretion is enhanced in rats^91,92. However, at high levels or long-term elevations, IL-1β acts as a chemoattractant and sets off a pro-inflammatory cascade in rats⁹³. The switch from low to high IL-1β levels has been hypothesized to be the factor that shifts individuals from pre-diabetes to T2DM⁹⁴. Using the mathematical model, bi-stable switch, the level of IL-1β was demonstrated to correlate with the stage of T2DM in humans⁹⁴. These findings strongly suggest that IL-1β has a key role in T2DM and possibly in disease progression.

The IL-1 receptor antagonist (IL-1RA) negatively regulates IL-1β. β-Cells can secrete IL-1RA, suggesting that they can regulate the islet microenvironment in humans^95,96. Many of the effects of IL-1β on glucose homeostasis are dependent on NF-κB and IL-1R signalling^88,89,93. Levels of IL-1RA are elevated before overt glucose intolerance occurs, which indicates a response to counter the pathologic effect of IL-1β^93,96. In individuals who go on to develop T2DM, circulating levels of IL-1RA are elevated⁹³ during the pre-diabetic state when patients have insulin resistance but β-cell compensation⁹⁷. In a subset of patients enrolled in the Whitehall II study, follow-up demonstrated that elevations in IL-1RA concentrations predicted subsequent development of T2DM⁹⁸. In mice fed a high-fat diet, the increase in IL-1RA levels precedes measures of glucose intolerance⁹⁶. These findings suggest that IL-1RA works to counter the elevated levels of IL-1β seen in the early pathology of T2DM.

The balance between IL-1RA and IL-1β suggests there is a role for IL-1β in islet homeostasis. At low levels, IL-1β promotes β-cell survival and glucose-stimulated insulin secretion in rats⁹². The role of low levels of IL-1β in β-cell function is particularly intriguing when considering IUGR (FIG. 5). Levels of IL-1β are increased in the liver of adult rat offspring born to mothers who received restricted food intake during pregnancy⁹⁹. Levels of IL-1β are also elevated in the placenta and amniotic fluid in IUGR rats in the bilateral uterine artery ligation model¹⁰⁰. Thus, it is possible that the lifelong high concentrations of IL-1β contribute to the impaired β-cell function that is often observed in offspring that experienced IUGR. It remains to be determined whether increases in levels of IL-1-β have a notable role in the pathogenesis of T2DM following IUGR, as it does in non-IUGR-induced T2DM.

Fig. 5 |. Pancreatic inflammation subsequent to IUGR.

Intrauterine growth restriction (IUGR) alters islet macrophage populations and inflammatory signals. Offspring affected by IUGR have elevated M2-driven inflammation, altered TGFβ signalling and increased levels of IL-1β in fetal and neonatal islets. Islets isolated from adults who were affected by IUGR have elevated levels of IL-1β and leptin.

TGFβ is key to islet development.

TGFβ is another key cytokine necessary for normal islet and pancreas development. TGFβ promotes the development of endocrine cells, particularly β-cells¹⁰¹. TGFβ is expressed in fetal islets and levels of expression vary during fetal development in mice¹⁰². One mechanism by which TGFβ influences islet development is via the activation of matrix metalloproteinase 2 (MMP2) at embryonic day 17–19. Additionally, the inhibition of MMP2 impairs islet morphogenesis in rats¹⁰³. The inhibition of TGFβ, via SMAD7, in the mouse fetus also negatively influences embryonic pancreas development, leading to β-cell hypoplasia¹⁰⁴.

Several studies in humans exposed to IUGR and animal models of IUGR have demonstrated that TGFβ signalling is aberrant in various tissues (FIG. 5). TGFα and TGFβ levels are elevated in cord blood collected from neonates exposed to IUGR compared with cord blood from those not exposed to IUGR¹⁰⁵. Using the elevated ambient temperature sheep model of IUGR, it was reported that 8% of the annotated genes had altered expression in fetal islets of offspring exposed to IUGR, and many of these genes were related to immune function⁶³. Islets of offspring exposed to IUGR also had a decreased expression of genes with protein products related to cell proliferation, which is consistent with decreased β-cell replication⁶³. In addition, altered TGFβ signalling was also found, suggesting that this might be one pathway that leads to decreased β-cell replication⁶³. How TGFβ contributes to the pathogenesis of IUGR-induced T2DM remains to be elucidated, but pathways implicated in the development of T2DM in adults might hold clues. TGFβ signalling promotes β-cell apoptosis and blunted glucose-stimulated insulin secretion, whereas inhibition of TGFβ signalling results in β-cell proliferation in β-cells isolated from adult humans¹⁰⁶. TGFβ signalling can be constitutively activated to directly test the role of TGFβ signalling, which resulted in increased β-cell apoptosis, decreased β-cell mass and, eventually, glucose intolerance¹⁰⁶. As β-cell death and blunted insulin secretion are key features of IUGR-induced T2DM, it is conceivable that TGFβ has an important role in the pathogenesis of T2DM in individuals who were growth-restricted at birth.

Insulin resistance in IUGR

Insulin resistance is a key component of IUGR-induced T2DM. While the scope of this Review focuses on immune dysregulation that results in β-cell dysfunction, immune pathways have also been implicated in the development of insulin resistance in humans and animal models of T2DM. We point the reader to two reviews for more information on the role of the immune system in insulin resistance^39,107. In the case of IUGR, insulin resistance is a key feature of the phenotype and is associated with increased basal insulin secretion, increased fasting levels of insulin and decreased glucose disposal^108,109. The administration of an IL-4-neutralizing antibody normalizes glucose disposal and basal insulin secretion⁵⁸, but, to date, no in-depth studies have been reported that dissect the underlying mechanisms.

Leptin in IUGR

Leptin is an adipokine that has a key role in the central control of bioenergetics and in the regulation of immunometabolism. The leptin receptor is expressed on many immune cells^110,111, and leptin regulates both innate and adaptive immune responses through modulation of immune cell survival and proliferation as well as by modulating the activity of immune cells¹¹². In innate immunity, leptin increases the cytotoxicity of natural killer cells and promotes the activation of granulocytes, macrophages and dendritic cells. Leptin also regulates macrophage M1-phenotype or M2-phenotype polarization¹¹². In the adaptive immune response, leptin promotes T cell survival¹¹³ by modulating the expression of anti-apoptotic proteins¹¹⁴.

Glucose disposal.

Storage of glucose as glycogen in tissues.

In general, leptin has pro-inflammatory properties and acts similarly to other acute phase reactants. Leptin upregulates the secretion of multiple inflammatory cytokines, including TNF, IL-6 and IL-12 (REF.¹¹⁵). By contrast, low levels of leptin, commonly observed in malnutrition, decrease T cell survival and proliferation as well as inflammatory cytokine production¹¹⁶. T cell glucose uptake and metabolism are also decreased when leptin levels are low¹¹⁶. Leptin levels are also low in fetuses affected by IUGR^117–122 and it is possible that this relative leptin deficiency might contribute to the immune dysfunction commonly observed in infants of low birth weight (FIG. 5). Interestingly, leptin levels increase with age in animal models of IUGR and humans exposed to IUGR, which correlates with increased adiposity and systemic inflammation^123–125. Furthermore, increased leptin levels in islets of IUGR animals correlates with increases in levels of pro-inflammatory cytokines⁴⁷. Whether or not leptin drives these changes in the immune profile of the IUGR islet remains to be determined.

Adiponectin in IUGR

Adiponectin is an adipokine that promotes anti-inflammatory activation of mouse macrophages, which leads to the production of cytokines such as IL-10, ARG1 and macrophage galactose N-acetylgalactosamine-specific lectin 1 (MGL1)¹²⁶. In addition, adiponectin results in a reduced production of pro-inflammatory cytokines such as MCP1, IL-6 and TNF¹²⁶. Adiponectin also inhibits the activation of many other immune cells¹²⁷

The effect of adiponectin on immune cells of offspring exposed to IUGR has yet to be studied; however, levels of adiponectin are positively correlated with birth weight in humans¹²⁸. Interestingly, in neonates exposed to IUGR that also had fetal velocimetry abnormalities, levels of adiponectin were statistically significantly lower than in neonates born SGA without fetal velocimetry abnormalities or in neonates born within the expected birth weight range for their gestational age¹²⁹. However, others have reported no difference in adiponectin levels when comparing SGA infants and infants within the expected birth weight range for their gestational age^122,130. Larger studies that are able to stratify neonates based on IUGR pathology as well as animal studies might clarify the effect of growth restriction on adiponectin levels.

Immunomodulation therapy

Immune-directed therapy in IUGR

Therapeutic approaches to prevent the development of T2DM in individuals who were affected by IUGR have primarily focused on nutritional interventions. The targeting immune pathways in IUGR has not been extensively studied; however, some studies report that this approach is beneficial. For instance, in our study (discussed previously), inflammation mediated by type 2 T helper cells is exaggerated in IUGR fetal rat islets⁵⁸. To test the pathological role of inflammation on the later development of T2DM, we blocked IL-4 with a neutralizing antibody injected on postnatal days 1–6. The IL-4-neutralizing antibody attenuated the IUGR-induced diabetic phenotype and restored β-cell function. These results suggest that the exaggerated type 2 T helper cell immune pathway is responsible for the subsequent development of β-cell dysfunction and might represent a potential therapeutic target.

In this same rat model of IUGR, in a separate study, we attenuated β-cell dysfunction through the administration of a glucagon-like peptide 1 (GLP1) agonist, exendin 4. This agonist was similarly injected on postnatal days 1–6, and adult offspring had improved β-cell function and did not develop T2DM¹³¹. We did not evaluate the immune system in these pups; however, GLP1 agonists are known to have immunomodulating effects. In addition, levels of pro-inflammatory mediators are reduced in adult patients treated with GLP1 agonists for T2DM^132,133. Therefore, it is possible that GLP1 agonists improved offspring β-cell function via immunomodulation.

In a pig model of IUGR, levels of the pro-inflammatory cytokines TNF and IL-10 were reported to be increased in the ileum of IUGR offspring⁵⁷. Elevations in the levels of these cytokines were associated with intestinal permeability and dysfunction⁵⁷. Feeding the piglets a diet rich in flax seed oil attenuated the cytokine response and improved intestinal permeability and function⁵⁷. Although this study did not address the development of T2DM, it is possible that similar therapeutic approaches might attenuate pancreatic inflammation.

T2DM and immunomodulating drugs.

Some drugs that are often used to treat hyperglycaemia during pregnancy might have immunomodulatory effects on the offspring. For instance, metformin is an anti-hyperglycaemic drug widely used for the treatment of insulin resistance and glucose intolerance associated with obesity and diabetes mellitus during pregnancy. Although no studies have examined the effects of maternal metformin treatment in IUGR pregnancies or on the immune function of the fetus and neonate, its general immunomodulatory properties make it an attractive candidate for the prevention of IUGR-induced T2DM. In non-cancer diseases, metformin generally acts to suppress pro-inflammatory pathways¹³⁴. Metformin also decreases the maturation of monocytes to macrophages¹³⁵ and promotes the alternative activation of macrophages (M2)^136,137. In obesity, metformin reduces levels of pro-inflammatory cytokines and chemokines in animal models^138,139 and in humans¹⁴⁰. Finally, metformin has been identified as an inhibitor of TGFβ as it prevents receptor binding and therefore downstream signalling in vitro¹⁴¹. Further investigation of the effect of metformin on offspring immune development, when used during pregnancy, is warranted.

Finally, several drug classes used for the treatment of T2DM have immunomodulatory effects. Most of these have not been used to treat pregnant women or children. Thus, their utility in preventing T2DM in the setting of IUGR remains to be determined. However, given that their mechanisms of action target immune pathways that are disrupted by IUGR, it is possible that they will have particular therapeutic value for the treatment of patients with this subclass of T2DM. For instance, thiazolidines bind to the nuclear receptor PPARγ and activate AMPK in rat skeletal muscle¹⁴². Both of these pathways are instrumental in immune activation and possibly explain the immunomodulating effects of thiazolidines. In addition, thiazolidines reduce the levels of pro-inflammatory mediators in patients with T2DM^143,144. Another class of drugs used to treat T2DM, dipeptidyl peptidase 4 inhibitors, reduces the levels of several pro-inflammatory proteins, including C-reactive protein, IL-6, sPLA2, ECAM, E-selectin and TNF^132–145. However, despite a comprehensive understanding of how these drugs alter the immune system, it is intriguing that there might be an underappreciated therapeutic advantage to manipulating the immune system in patients with T2DM.

Conclusions

Immune pathways are altered early in the pathogenesis of IUGR-induced T2DM, and this change precedes β-cell dysfunction. Studies have shown that the prevention of inflammation in neonates affected by IUGR normalizes β-cell function. These findings provide strong evidence that altered immune system function has a central role in the pathogenesis of T2DM in offspring affected by IUGR. However, additional studies are needed using other species and models to demonstrate the importance of altered immune system function in this process. Numerous studies have demonstrated that islet macrophages are critically important for normal islet homeostasis. However, we do not yet know whether the macrophage is the immune cell type that is directly responsible for the abnormal islet phenotype in animal models of IUGR or whether the critically important crosstalk between different immune cell populations is altered in animal models of IUGR. Finally, human studies are lacking and are necessary to identify therapeutic targets that can be directed to IUGR-induced T2DM.

Key points.

• Fetal immune development is susceptible to an abnormal intrauterine milieu, and alterations have been implicated in the development of type 2 diabetes mellitus (T2DM) following intrauterine growth restriction (IUGR).

• Pancreatic islet macrophages are instrumental in islet development and homeostasis in adults.

• Levels of cytokines and immune mediators that are involved in T2DM pathogenesis are also elevated in offspring exposed to IUGR.

• Leptin stimulates IL-1β production in islets, and levels of leptin are reduced in fetal islets and elevated in adult islets of offspring exposed to IUGR.

• The results of limited reports evaluating the therapeutic effect of immunomodulation are promising for the treatment of IUGR-induced T2DM.