Central actions of insulin during pregnancy and lactation

Abstract

Pregnancy and lactation are highly metabolically demanding states. Maternal glucose is a key fuel source for the growth and development of the fetus, as well as for the production of milk during lactation. Hence, the maternal body undergoes major adaptations in the systems regulating glucose homeostasis to cope with the increased demand for glucose. As part of these changes, insulin levels are elevated during pregnancy and lower in lactation. The increased insulin secretion during pregnancy plays a vital role in the periphery; however, the potential effects of increased insulin action in the brain have not been widely investigated. In this review, we consider the impact of pregnancy on brain access and brain levels of insulin. Moreover, we explore the hypothesis that pregnancy is associated with site-specific central insulin resistance that is adaptive, allowing for the increases in peripheral insulin secretion without the consequences of increased central and peripheral insulin functions, such as to stimulate glucose uptake into maternal tissues or to inhibit food intake. Conversely, the loss of central insulin actions may impair other functions, such as insulin control of the autonomic nervous system. The potential role of low insulin in facilitating adaptive responses to lactation, such as hyperphagia and suppression of reproductive function, are also discussed. We end the review with a list of key research questions requiring resolution.

1 ∣. PREGNANCY

1.1 ∣. Glucose homeostasis undergoes significant changes during pregnancy

Glucose is the main energy source used by the fetus. Therefore, the mother must provide a constant supply across the placenta for the successful growth and development of offspring. Glucose transport from mother to fetus occurs by passive diffusion, relying on the maintenance of a higher glucose concentration on the maternal side compared to the fetal side. As pregnancy progresses and the fetus grows, this glucose gradient becomes harder to maintain and requires adaptations in maternal glucose regulation.

Insulin is the key hormone regulating blood glucose levels. When glucose levels rise, insulin is secreted to increase glucose uptake into peripheral insulin-sensitive tissues. As part of the normal response to pregnancy, maternal tissues such as muscle and fat become relatively insulin resistant^1-5 thereby impeding glucose uptake (Figure 1) and favouring increased delivery of glucose to the fetus. These maternal tissues instead increase use of other fuel sources such as fatty acids and ketone bodies.⁶ However, excessive exposure to glucose can also be detrimental to optimal fetal growth. Therefore, part of the changes in glucose regulation act to prevent excessive glucose exposure. For example, normal pregnancy reduces the threshold for glucose-stimulated insulin secretion from the insulin-secreting beta-cells of the pancreas,⁷ so that, in response to increases in maternal blood glucose, there is a greater insulin response to attempt to rapidly reduce the high blood glucose level despite the relative insulin resistance of many maternal tissues (Figure 1). Alongside this lowered threshold for glucose-stimulated insulin secretion, beta-cells undergo expansion during pregnancy, and this expansion begins prior to the development of peripheral insulin resistance in other tissues.⁷ The pregnancy-induced beta-cell expansion represents an “anticipatory” adaptation, a physiological change that meets the future needs of increased insulin during pregnancy.⁸ Activation of the prolactin receptor, either through the hormone prolactin or its homologue, placental lactogen, plays a key role in driving beta-cell expansion during pregnancy.^7-9

FIGURE 1.

The key question of central insulin action in the brain during pregnancy: does elevated insulin level increase central insulin function or does central insulin resistance develop? Insulin is secreted from beta-cells in the pancreas (as indicated by the brown immunoreactive staining in the cross-section of pancreas) and maternal adaptation in glucose regulation leads to increased insulin levels to compensate for increased peripheral insulin resistance during pregnancy. The effect of increased insulin levels during pregnancy on central insulin functions are yet to be thoroughly explored

The fine balancing act to maintain optimal levels of glucose transfer to the fetus and to protect against excessive exposure tends to increase circulating insulin levels during pregnancy. Glucose-stimulated insulin secretion is elevated^7,8 and, in humans, fasting insulin levels increase.^10-12 Whether similar rises in circulating insulin levels occur in pregnant animal models depends on the species, satiety status and experimental protocols. In the often-studied rat, under fasting conditions, pregnant insulin levels are similar to virgin levels.^13-15 Pentobarbital anaesthetised pregnant rats also have similar insulin levels as virgins.^16,17 Similar to many anaesthetics, pentobarbital can alter the regulation of glucose^18-20 and thus caution should be taken in the interpretation of insulin levels measured under anaesthetic. However, unanaesthetised rats, mice and rabbits all show elevated basal insulin levels in the non-fasting state.^15,21-24 Our data in rats and mice indicate that this elevation is detectable from mid-pregnancy,^21,23 although others have not seen significant increases until late pregnancy (day 18 in rats).²⁵ Some studies have only investigated a single time point in late pregnancy, showing increased insulin levels compared to virgins following 6 hours of food deprivation (although the timing of this fast was not stated).²⁶

In the periphery, insulin plays the key role with respect to promoting uptake of glucose by insulin-sensitive tissues, such as liver, muscle and fat. Insulin receptors are also found in the brain ^27,28; however, insulin does not drive the uptake of glucose by the brain as it does for peripheral tissues. Nevertheless, insulin can act in the brain to regulate a wide range of other functions, including food intake and energy homeostasis, peripheral glucose levels, reproduction, the sympathetic nervous system and cognition.²⁹ The increase in circulating insulin levels during pregnancy therefore would suggest the potential for increased insulin action in the brain during this state. However, the actions of insulin in the brain also depend on the rate of uptake of insulin into the brain and insulin receptor responsiveness. Indeed, although increases in peripheral insulin are required for optimal offspring development, increases in some of the central functions of insulin, in particular inhibition of food intake and suppression of hepatic glucose production, would be counterproductive. Therefore, the aim of this review is to discuss the impact of pregnancy on brain access and responsiveness to insulin (Figure 1), as well as the resulting changes in select central insulin actions.

1.2 ∣. Insulin access to the brain during pregnancy

Insulin is produced in the periphery and to have effects in the brain, it needs to gain access across the blood-brain barrier (BBB).^30,31 Insulin is too large to simply diffuse into the brain but, instead, moves across the BBB via an active, saturable trans-endothelial transport mechanism. In rodent models of obesity causing insulin resistance and hyperinsulinaemia, impaired access of insulin into the brain is considered to play a major role in the attenuated central insulin actions.^32,33 During pregnancy, insulin transport into the brain is mostly unaffected, although transport of insulin across the choroid plexus, into cerebrospinal fluid (CSF), is increased.¹⁷ Thus, although other insulin-resistant states such as obesity and ageing decrease insulin transport into brain,^32,34,35 this does not appear to be the case during pregnancy.

The level of insulin in the brain is affected not only by plasma insulin levels and transport into brain, but also by degradation. Interestingly, parenchymal degradation of insulin in the brain is elevated during pregnancy, which may explain why pregnancy does not alter brain insulin levels, despite elevated plasma levels.¹¹ In vitro studies have shown that insulin itself can increase levels of insulin-degrading enzyme (IDE) in neurones³⁶ and others have shown that oestrogen can increase IDE in primary cultures of rat hippocampal neurones.³⁷ Whether elevated degradation of insulin in the brain is a result of a pregnancy-induced increase in IDE level or activity, as well as what role hormonal changes associated with pregnancy play in driving any potential change in degradation, remains to be investigated. Another outstanding question is what is the specific location of this increased degradation?

The impact of pregnancy on CSF levels of insulin has also been investigated, with varying results. Insulin levels were decreased in both late pregnancy in rats³⁸ and rabbits,¹⁶ but, in another study, they were unchanged in late pregnant rats.¹⁷ The variability may be a result of methodological differences or, more simply, the action of pregnancy to increase plasma insulin levels and transport into CSF on the one hand but increase degradation on the other. Regardless, insulin transport via the CSF is not considered to be a major access point of insulin to the brain, and so it is unclear what these CSF values reflect in terms of functionality.^17,30 Overall, the evidence suggests that pregnancy does not significantly alter levels of insulin in brain interstitial fluid.

1.3 ∣. Insulin-induced intracellular signalling in the brain during pregnancy

During pregnancy, in tissues such as skeletal muscle and adipose tissue, insulin resistance has been associated with attenuated insulin-induced activation of intracellular signalling pathways. Insulin predominantly acts through a phosphoinositide 3-kinase (PI3K) pathway,³⁹ and pregnancy-induced insulin resistance in skeletal muscle down-regulates various steps in this signalling cascade, including reduced phosphorylation of the insulin receptor (IR) and reduced expression of insulin receptor substrate 1 (IRS-1).⁴⁰ Emerging evidence also suggests the appearance of central insulin resistance (Figure 2). For example, early in the dark phase, which is a time of increased food consumption in rats, plasma insulin concentrations are greater in pregnant rats (day 14 of gestation) than in virgins. Despite this, pregnant rats show markedly lower phosphorylated Akt (pAkt) levels in the arcuate nucleus, a key site of insulin action in the brain, compared to virgin rats.¹⁵ Even in other areas of the hypothalamus, pAkt levels were similar between pregnant and virgin rats despite the difference in plasma insulin.¹⁵ Of course, Akt is not solely phosphorylated because of interactions between insulin and its receptor, and so these data are only suggestive of impaired sensitivity of neurones to endogenous insulin. When insulin-induced pAkt was directly assessed following acute, central administration of exogenous insulin, insulin sensitivity in the arcuate and ventromedial nuclei of the hypothalamus (VMH) was substantially attenuated during pregnancy. Yet other hypothalamic areas, the paraventricular nucleus (PVN) and dorsomedial nucleus (DMH), remained responsive in terms of insulin-induced pAkt.¹⁵ These data indicate that, within the hypothalamus, there are site-specific reductions in insulin sensitivity during pregnancy (Figure 2).

FIGURE 2.

Changes in cellular insulin signalling in various hypothalamic nuclei in the non-pregnant and pregnant state. Insulin is transported from the blood into the brain, where it acts in various regions, including hypothalamic areas such as the arcuate nucleus, paraventricular nucleus (PVN), dorsomedial nucleus (DMH) and ventromedial nucleus (VMH). Interaction of insulin with its receptor leads to phosphorylation of Akt (pAkt) as part of insulin-induced intracellular signalling cascade. In the arcuate nucleus, insulin suppresses (solid red arrow) mRNA expression of orexigenic neuropeptide Y (NPY) and agouti-related peptide (AGRP) and stimulates (solid green arrow) mRNA expression of anorectic pro-opiomelanocortin (POMC). Phosphatase and tensin homolog (PTEN) is a negative regulator of Akt and phosphorylation leads to deactivation of this function of PTEN. During pregnancy (middle) insulin transport into the brain is not reduced, whereas insulin-induced pAkt in the arcuate nucleus and VMH is attenuated. NPY, AgRP and POMC mRNA expression in the arcuate nucleus is dissociated from peripheral insulin levels. Reduced phosphorylated PTEN (pPTEN) in the VMH suggests that a reduction in the deactivation of this inhibitory signalling molecule might contribute to attenuated insulin-induced pAkt during pregnancy. During lactation (bottom), the low plasma insulin concentrations are likely to limit the effects of insulin in the brain. Exogenously administrated insulin can regulate NPY, AgRP and POMC mRNA levels, indicating that there is a restoration of insulin sensitivity in these neurones compared to pregnancy, yet the low endogenous insulin levels likely contribute to high NPY and AgRP mRNAs and low POMC mRNA observed during lactation. Whether there are changes in insulin transport into the brain or insulin responses in other neurone populations during lactation has yet to be determined. Green arrows indicate stimulatory effects of insulin, whereas red arrows indicate inhibitory effects. Grey arrows/text indicates the attenuated actions of insulin during pregnancy. 3V, third ventricle

Within the arcuate nucleus, insulin has been shown to act on key neuronal populations involved in the regulation of energy homeostasis: the anorectic pro-opiomelanocortin (POMC) neurones and orexigenic neuropeptide Y (NPY)/agouti-related peptide (AgRP) neurones. In alignment with its anorectic function, insulin can increase the expression of POMC mRNA⁴¹ and inhibit the expression of NPY mRNA and AgRP mRNA in the arcuate nucleus^42-44 (Figure 2). During pregnancy, POMC mRNA is relatively constant or reduced,^45-49 and NPY mRNA and AgRP mRNA are either similarly constant or increased.^46,48-50 Although other factors can influence expression of these neuropeptides, these changes are opposite to those predicted based on the increases in insulin levels, hinting further that insulin sensitivity is reduced in the arcuate nucleus in particular at this time.

As yet, the mechanisms underlying the suppressed insulin-induced pAkt in the arcuate nucleus and VMH during pregnancy are unknown. IR protein levels in these two nuclei do not change during pregnancy (Figure 2), ruling out a reduction in IR expression as a cause of the attenuated insulin sensitivity.¹⁵ In animal models of obesity, arcuate IR expression is also unchanged,^51,52 and central insulin resistance is instead associated with impaired insulin signalling as a result of increases in tyrosine phosphatases that dephosphorylate the IR and IRS.⁵³ Further investigations are required to determine if similar mechanisms are involved during pregnancy. In the VMH, reduced levels of phosphorylated phosphatase and tensin homolog (PTEN) during pregnancy have been reported, potentially leading to increased activity of this negative regulator of PI3K,¹⁵ which might contribute to reduced insulin signalling (Figure 2). Interestingly, pPTEN levels were unaffected during pregnancy in the arcuate nucleus, suggesting that different mechanisms underlie changes in insulin sensitivity in these two regions that show this pregnancy-associated reduction in insulin-induced pAkt.¹⁵

1.4 ∣. Impact of pregnancy on the central functions of insulin

As introduced earlier, IR is expressed in many brain regions,^27,28 particularly the hypothalamus, and insulin influences a wide range of brain functions. Region-specific “insulin resistance” in the brain of pregnant animals, as described in the previous section (Figure 2), would suggest that some of these actions may be unaffected, whereas others might be impaired. Insulin resistance in the arcuate nucleus and VMH may be adaptive to prevent insulin actions that may hinder optimal pregnancy outcomes, at the same time as allowing insulin actions that remain responsive to help meet the demands of pregnancy. Although many of the roles of insulin in the brain have not been well characterised during pregnancy, perhaps the best studied are the involvement of insulin in the central regulation of energy balance and in the autonomic control of cardiovascular function.

1.4.1 ∣. Increases in food intake during pregnancy

There is extensive evidence indicating that insulin can act in the brain, specifically the hypothalamus, to influence food intake and energy homeostasis. IR is found in key hypothalamic areas involved in energy regulation, such as the arcuate nucleus, VMH and PVN. Early investigations in baboons demonstrated that central administration of insulin can suppress food intake⁵⁴ and this has subsequently been repeated in many other species, including rats^55-57 and mice.⁵⁸ Central infusions of insulin antibodies or IR antisense oligonucleotides to block central insulin action drive an increase in food intake and bodyweight.^57,59,60 There is not complete uniformity in results with respect to investigating the anorectic effects of central insulin as some studies report no satiety effect of insulin with only small changes in protocols.^61,62

Pregnancy is associated with increased food intake despite elevations in insulin levels, which is highly suggestive that the brain becomes insensitive to its anorectic actions. This has yet to be confirmed with a functional assessment of food intake following insulin administration during pregnancy. Given that the arcuate nucleus is a key site for insulin regulation of food intake and that insulin-induced intracellular signalling is impaired in this area during pregnancy,¹⁵ it would certainly be predicted that there is a lack of central insulin-induced satiety during pregnancy. One of the mechanisms by which central insulin action is thought to suppress food intake is through its ability to increase sensitivity to acute satiety factors such as the meal terminating hormone cholecystokinin (CCK).⁶³ Pregnancy in rodents increases meal duration⁶⁴ and impairs the acute satiety actions of CCK.⁶⁵ It is possible that any attenuation of central insulin action may contribute to this reduction in CCK sensitivity during pregnancy, although this has yet to be investigated directly. Similar to central pregnancy-induced leptin insensitivity,^66,67 it also remains unknown whether insulin insensitivity in brain circuits regulating food intake drives hyperphagia during pregnancy, or merely facilitates it by removing potential inhibition via this hormone.

1.4.2 ∣. Reductions in physical activity during pregnancy

Insulin action in the brain has also been suggested to influence levels of physical activity, although this effect can be variable depending on the physiological situation. Central insulin administration can increase locomotor activity in lean mice but diet-induced obese mice are insensitive to this effect of insulin.⁶⁸ Other studies have shown that a high dose of insulin given centrally can suppress locomotor activity⁶⁹ and that insulin activation of melanin-concentrating hormone (MCH) neurones in the lateral hypothalamus leads to a suppression of locomotion.⁷⁰ The insulin-induced decrease in locomotion through MCH neurones is specifically activated when exogenous insulin is elevated, such as in diet-induced obesity, and may represent a neuronal population that does not develop insulin insensitivity in obesity.⁷⁰ Mice with a deletion of insulin receptors from either NPY, AgRP or POMC neurones show no physical activity phenotype, suggesting these key insulin-responsive neuronal populations are unlikely to mediate any effect of insulin on locomotion.^71,72 Pregnancy is associated with reductions in physical activity^64,73-75 and whether the elevated insulin levels contribute to these lower levels of ambulation through activation of MCH neurones⁷⁰ or through insensitivity of other neuronal populations that mediate the insulin-induced increases in physical activity⁶⁸ is yet to be determined, although it poses an interesting unresolved question.

1.4.3 ∣. Pregnancy-induced changes in cardiovascular regulation

The maternal cardiovascular system undergoes major adaptations to support the developing fetus. These include a primary decrease in systemic vascular resistance, which drives, at least in part, substantial increases in heart rate (HR), cardiac output and blood volume, all of which assure adequate delivery of oxygen and nutrients to the uteroplacental unit. Pregnancy in humans also progressively increases muscle sympathetic nerve activity (SNA); in animal models (mostly rodent) increases in basal SNA to several organs have been reported.⁷⁶ This sympathoexcitation helps to counteract widespread vasodilation, thereby limiting the fall in arterial pressure, and also promotes fluid retention. However, the mechanisms have yet to be fully explained. Conversely, the sensitivity of the baroreceptor reflex progressively decreases during pregnancy.⁷⁶ Baroreflex control of both HR and lumbar or muscle sympathetic activity are impaired, which underlies the increased incidence of orthostatic hypotension and a reduced ability to maintain blood pressure during haemorrhage.^77,78

Similar to pregnancy, insulin increases SNA to several organs, although lumbar SNA (in rats) and muscle SNA (in humans) are particularly stimulated. Insulin binds to receptors in only one brain site, the arcuate nucleus, to increase SNA^79-81 through inhibition of sympathoinhibitory NPY neurones and activation of sympathoexcitatory POMC neurones that project to the hypothalamic paraventricular nucleus (PVN).^82,83 Glutamatergic neurones in the PVN are stimulated, which drive vasomotor neurones in the rostral ventrolateral medulla.⁸⁴ Also, in both rodents and humans, acute infusion of insulin, either central administration or i.v. administration, respectively, enhances baroreflex sensitivity.^85,86

The pregnancy-induced increase in basal SNA is driven in part by the arcuate nucleus, specifically by decreasing NPY and increasing POMC inputs to the PVN, just like insulin action⁸⁷; therefore, insulin was considered a plausible mediator. However, in late pregnant rats, the sympathoexcitatory response to i.c.v. insulin (or leptin) was completely abolished.¹⁷ More importantly, bilateral local blockade of arcuate IR, through nano-injections of the IR antagonist S691, failed to lower SNA,¹⁷ unlike its action in insulin-resistant obese male rats.⁸⁸ Therefore, insulin does not contribute to the basal sympathoexcitation of pregnancy.¹⁷

As noted above, during pregnancy, baroreflex-induced increases in SNA in response to hypotension are impaired. This change may be partially a result of the marked elevations in basal SNA, closer to the highest SNA levels reached when blood pressure is low, and also because of greater inhibition of vasomotor neurones in the rostral ventrolateral medulla.⁷⁶ Because insulin does not contribute to basal sympathoexcitation, as a result of the development of stark central insulin resistance, insulin also cannot explain the increase in basal SNA closer to the baroreflex maximum. The cardiac or HR baroreflex is also impaired; HR baroreflex sensitivity or gain is reduced and, similar to SNA, maximal HR levels in response to hypotension are decreased during pregnancy. These changes are a result of suppressed responsiveness of both the sympathetic and parasympathetic control of HR.⁷⁵

It was hypothesised that whole-body insulin resistance may contribute to the mechanisms underlying arterial cardiac baroreflex dysfunction during pregnancy.⁸⁹ In support of this concept, other conditions show similar links between whole-body insulin resistance and baroreflex dysfunction.⁹⁰ In pregnant rabbits, reductions in whole-body insulin sensitivity were correlated with reductions in baroreflex gain, demonstrating a similar time course of progression.¹⁶ Similarly, in rats, both baroreflex sensitivity and whole-body insulin sensitivity are reduced at a similar time in mid- and late pregnancy, with a transient resumption of near-normal baroreflex and insulin sensitivities between these time points.^25,89 Although it was initially hypothesised that peripheral insulin resistance might reduce access of insulin to the brain,¹⁶ as outlined above, recent work showing that pregnancy does not impair insulin transport into the brain or reduce brain insulin levels¹⁷ does not support this initial hypothesis. Nevertheless, i.c.v. insulin infusion in conscious³⁸ and urethane-anaesthetised¹⁷ rats did improve cardiac baroreflex gain or sensitivity without normalising the decrease in maximal HR achieved with hypotension. Thus, it appears that this depressed feature of baroreflex function (gain) retains some sensitivity to insulin, whereas the blunted maximal reflex tachycardia could be secondary to the central insulin resistance that develops. In support, i.c.v. infusion of the artificial CSF vehicle decreased the HR baroreflex maximum in conscious virgin but not pregnant rats, suggesting a factor in brain that normally supports baroreflex function in virgin (but not pregnant) rats was diluted by the infusion. As previously discussed,³⁸ this factor could be insulin (levels or responsiveness). Alternatively, a factor that contributes to peripheral insulin resistance could in parallel decrease hypothalamic insulin sensitivity, thereby causing the marked decreases in HR baroreflex function that occur during pregnancy. Thus, the links between the reduced (peripheral and central) insulin sensitivity and the substantial impairment of cardiac baroreflex function during pregnancy remain unresolved and require further study.

1.4.4 ∣. Hepatic glucose production during pregnancy

Hepatic glucose production plays an important role in maintenance of blood glucose levels (particularly when food is scarce) and is readily responsive to changes in energy balance. Insulin directly suppresses hepatic glucose production. Central insulin action can also regulate blood glucose levels,⁹¹ particularly through the control of hepatic glucose production.^91-94 Arcuate nucleus AgRP neurones are a key site of action, since mice with a specific deletion of insulin receptors from arcuate AgRP neurones show increased hepatic glucose production.⁴² Intranasal insulin administration in humans can also suppress glucose production independently of changes of peripheral insulin levels, indicating that preclinical animal studies provide a valuable model for central insulin's ability to inhibit hepatic glucose production.⁹⁵

Hepatic glucose production is increased in later stages of pregnancy in humans ^96-98 and rats.^14,99 As described above, the central actions of insulin on hepatic glucose output appear to be mediated by AgRP neurones. Because insulin-induced pAkt is reduced in the arcuate nucleus during pregnancy,¹⁵ reduced insulin sensitivity of AgRP neurones may underlie the increase in glucose production. However, this has yet to be directly examined. Hepatic glucose production is complex and there are multiple redundancies in the control of this process⁹⁴ that could allow for other mechanisms to play a more critical role in driving increases in hepatic glucose production during pregnancy, such as decreased liver sensitivity to insulin.¹⁴ Nevertheless, attenuated central insulin control of hepatic glucose production during pregnancy is an intriguing hypothesis that warrants investigation.

1.4.5 ∣. Cognition

Cognition, particularly memory and learning, can be affected by pregnancy. The anecdotal concept of “pregnancy brain” or “mommy brain” is certainly widely accepted and a recent meta-analysis of studies focusing on objective measures of cognition found a significant decline in cognitive function in pregnant women compared to non-pregnant women, particularly in the third trimester.¹⁰⁰ Memory recall was the main function that was negatively impacted.^100-102 In self-reported studies, a wider range of cognitive issues were impaired during pregnancy, including an inability to concentrate or maintain attention, increased confusion and a lack of coordination.^103-105

Insulin influences cognition, with acute increases in insulin proving to be beneficial in both human and animal studies, in particular, by improving memory formation.¹⁰⁶ This action likely is initiated in the hippocampus, where insulin influences signalling cascades that underlie hippocampal plasticity, learning and memory.^107-109 Furthermore, long-term studies using intranasal insulin administration in humans have shown improved memory without any negative effects on glycaemia.¹¹⁰

Insulin resistance in the brain has been linked as a contributing factor to cognitive decline in Alzheimer's disease, ageing, obesity and type 2 diabetes.¹¹¹ Given that a decline in cognitive abilities can be associated with states of systemic insulin resistance when insulin levels are typically elevated, it is suggested that central insulin resistance may also develop. Insulin resistance in the hippocampus could lead to reductions in insulin's effects on cognition despite elevated insulin levels.¹¹¹ Directly increasing brain insulin levels can overcome this reduction in central insulin sensitivity because models of impaired cognition and insulin resistance can show improvements in cognition with chronic intranasal administration of insulin.^112,113 This suggests that the mechanism underlying this central insulin insensitivity in these cases is more likely to involve access and the availability of insulin in the brain rather than impaired intracellular activation. Given that insulin transport into the brain is unaffected by pregnancy,¹⁷ whether a mechanism similar to this develops in pregnancy appears unlikely, although future work will be needed to confirm this. It is tempting to speculate that decreased insulin sensitivity in the hippocampus may contribute to any transient deterioration in cognitive function experienced during pregnancy. This is an unexplored area and perhaps initial work should focus on whether there is reduced insulin sensitivity in the hippocampus similar to the arcuate nucleus and VMH in the hypothalamus during pregnancy.

Of note, the results from animal studies are not indicative of declined cognitive function during pregnancy,¹¹⁴ raising the question of whether pregnancy may represent more of a shift in focus of aspects of cognition with respect to optimising tasks important for nurturing offspring that are not adequately tested for in current measures of cognitive function.¹¹⁵ Such an adaptive view of “mommy brain” as opposed to the maladaptive perception that is commonly considered in the popular media, is more consistent with the evolutionary perspective that hormone-induced adaptations in the maternal brain are predominantly beneficial to ensure a healthy pregnancy.¹¹⁶

1.4.6 ∣. Other actions of insulin in brain

Central insulin action has been implicated in many other functions, including increases in body temperature via actions in the preoptic area of the hypothalamus^117,118 and changes in olfaction via action in the olfactory bulb.^119,120 However, as with most of the specific actions of central insulin discussed above, there is a lack of studies directly assessing these effects in pregnancy. Thus, although the growing list of central actions of insulin currently provides many opportunities for speculation about how this might be impacted by pregnancy, it also means that there are many opportunities for future investigation.

2 ∣. LACTATION

2.1 ∣. Glucose homeostasis during lactation

Lactation is an enormous metabolic investment for the mother, and in species with multiple young, such as rodents, the energy demands of milk production are extreme. Lactating rats and mice consume two to three times their virgin level of food^121,122 to meet these energy demands. Glucose is a key requirement for milk production, especially for the production of lactose in milk. As a result, blood glucose levels are on the lower side of normal as a result of avid sequestration by the mammary gland and, in parallel, insulin levels also decrease. Lower blood glucose levels presented to the pancreas, rather than changes in beta-cell function, are thought to mediate reduced insulin secretion during lactation.^123,124 Despite the relative hypoinsulinaemia,^125,126 selective resistance is maintained in some maternal tissues, such as fat and muscle, to direct glucose to the mammary gland.^127,128 Over the years, the hypoinsulinaemia of lactation has been hypothesised to contribute to a number of functional changes in brain, in particular hyperphagia and infertility.

2.2 ∣. Impact of lactation on the central functions of insulin

2.2.1 ∣. Hyperphagia during lactation

Although the energetic drain of lactation in terms of milk production might appear to be a sufficient driving factor for hyperphagia, the suckling stimulus and associated hormonal changes alone can drive higher levels of food intake compared to virgin levels.¹²² Along with increased food intake, there is increased NPY mRNA and AgRP mRNA in the arcuate nucleus during lactation.^129-131 Furthermore, during lactation, NPY mRNA expression greatly increases in the DMH, contributing to the hyperphagia of this state.^132,133 It is possible that the low levels of insulin during lactation facilitate these increases in orexigenic neuropeptide mRNA and thus contribute to the hyperphagia of lactation by removing insulin inhibition on these neurones. However, chronic infusion of insulin during lactation in rats to increase insulin levels to control virgin levels has no effect on food intake,¹²⁶ suggesting a level of insulin resistance. Interestingly, these infusions did reduce the high NPY mRNA levels seen in lactation down to virgin levels, and slightly but significantly reduced the lactation-induced increase in AgRP mRNA.¹²⁶ The low levels of POMC mRNA during lactation were also elevated after this insulin treatment.¹²⁶ First, these data suggest that the arcuate nucleus remains somewhat sensitive to changes in insulin during lactation, as reflected by the ability of insulin to regulate NPY, AgRP and POMC mRNAs. Second, the data suggest that, at this stage of lactation, these arcuate nucleus orexigenic peptides are not the major factors inducing the hyperphagia of lactation.

The ability of insulin to regulate NPY, AgRP and POMC mRNAs¹²⁶ would suggest that insulin signalling, or at least insulin-induced pAkt, in the arcuate nucleus is restored at some point between mid-pregnancy and mid-lactation, although this has yet to be specifically investigated. Furthermore, whether insulin sensitivity is restored during lactation in the VMH, an area where attenuated insulin-induced pAkt signalling is associated with pregnancy,¹⁵ remains unknown. Further investigations are required to determine when this transition in central insulin sensitivity between pregnancy and lactation takes place, and whether there are differences as to if or when this process occurs in various different insulin-responsive regions.

2.2.2 ∣. Lactational infertility

Lactation is a state of suppressed reproductive function, characterised by a suppression of pulsatile luteinising hormone (LH) secretion, which is a key part of the reproductive axis driving follicle development and ovulation. There is a long history of work linking metabolic fuel availability with reproductive ability¹³⁴ and hence one hypothesis for the suppression of LH pulses during lactational infertility is that this state of negative energy balance, and the hormonal milieu that goes along with this, contribute to this suppression. Increases in insulin have been associated with resumption of LH pulses following food restriction,¹³⁵ and increasing insulin levels can lead to increased LH secretion.¹³⁶ Furthermore, female mice lacking IR in the brain demonstrate lower LH levels,¹³⁷ although this should be interpreted with caution because one-off LH measurements are not a reliable substitution for serial measurements of pulsatile LH secretion. In terms of lactation, however, the evidence would suggest it is unlikely that low insulin levels act as a metabolic signal to the brain that suppresses fertility at this time. Studies utilising continuous infusion of insulin during lactation to increase insulin to virgin levels have not led to restoration of LH secretion,^126,138 indicating that low central actions of insulin alone do not drive suppression of LH as part of lactational infertility.

2.2.3 ∣. Functional changes in magnocellular neurones during lactation

Magnocellular neurones (MCNs) in the supraoptic nucleus of the hypothalamus have been proposed to act as metabolic sensors by utilising insulin-dependent mechanisms for glucose uptake, similar to VMH glucose-sensing neurones. More specifically, glucose depolarises the membrane through glucose kinase (GK)-mediated inaction of ATP-sensitive potassium channels, which in turn stimulates oxytocin (OT) and vasopressin (VP) secretion from the posterior pituitary.¹³⁹ During lactation, blocking GK activity in these neurones does not prevent the insulin and glucose-stimulated release of OT and VP as it does in males (nonlactating females were not included as a control), indicating perhaps a change in intracellular signal transduction by which these neurones respond to insulin and glucose with release of OT and VP.¹⁴⁰ The exact mechanism has not been defined, although it may be related to a potential change in how these neurones take up and use glucose during lactation, switching from low-affinity GK, which is used as a metabolic sensor, to higher affinity glycolytic hexokinases to provide increased levels of fuel for cellular activity.¹⁴⁰ Thus, during lactation, MCN neurones maximise their glucose uptake and metabolism to meet the demands of increased OT production required for milk letdown, at the expense of their ability to act as a glucose monitor. In parallel, astrocytes withdraw support for MCN, allowing for more synaptic connections between MCNs^141,142 and more efficient and synchronised activity in OT release.¹⁴³ Thus, it is possible that, during lactation, MCNs move from lactate for cellular energy to other fuel sources such as insulin-mediated glucose uptake.¹⁴⁰ Although much of the details of this proposed change in metabolic process of MCN during lactation require additional work for confirmation, this does provide an example of how one specific neuronal population may undergo a change in insulin-mediated function during lactation.

3 ∣. BEYOND PREGNANCY AND LACTATION

The process of pregnancy and lactation represent a major biological challenge to the female body and, as such, is associated with a wide range of physiological adaptations to meet this challenge. In particular, becoming a mother leads to many alterations in the brain that can have both short- and long-term impacts.¹⁴⁴ Whether the decreases in central insulin sensitivity that develop during pregnancy,¹⁴⁵ despite some restoration in lactation,¹²⁶ have any lasting impact on insulin function in the maternal brain has yet to be investigated; however, there are some intriguing possibilities.

Recent work has highlighted that undergoing pregnancy and lactation (termed ‘reproductive experience’) leads to long-term biological changes in body weight regulation, in particular a reduction in locomotion activity.^145,146 Could a long-term change in insulin sensitivity in the lateral hypothalamus contribute to this lower locomotion activity? As described above, the exact role of insulin with respect to promoting or suppressing locomotion in different physiological states is not completely understood, although it begs the question of what this effect may be in reproductively experienced female.

Evidence for long-term cognitive changes as a result of parity in humans is somewhat equivocal; however, in animal models, it has been demonstrated that reproductive experience leads to enhanced cognitive function, particularly with memory and learning (reviewed in 144). Alongside this, reproductive experience in animal models induces positive neuroplasticity, such as increases in synaptic proteins,^147,148 a higher level of hippocampal neurogenesis^149-151 and a general slowing down of age-dependent decline in neurogenesis.¹⁴⁴ Insulin has been shown to promote neuroplasticity both in development and in the adult, contributing to a healthy brain.¹⁵² Future investigations examining long-term changes in insulin sensitivity in the hippocampus after pregnancy and lactation may reveal a contributing role for insulin in these changes in neuroplasticity and cognition after reproductive experience.

4 ∣. PERSPECTIVES/CONCLUSIONS

Pregnancy and lactation are times of major adaptation in the regulation of glucose homeostasis as the maternal body adjusts to meet the metabolically demanding challenge of fetal growth and development, as well as subsequent milk production for offspring following birth. Peripheral insulin resistance develops during pregnancy, and recent studies indicate that central insulin resistance is also present, leading to reduced insulin actions in the brain. This central insulin resistance is site-specific and may reflect adaptive changes in the brain that decrease some of insulin's central effects at the same time as maintaining other functions that support a healthy pregnancy.

Yet, our understanding of these homeostatic mechanisms is limited. Much more information is required, especially given the importance of glucose and insulin levels and their actions on fetal development, as well as maternal health. We offer the following as key unanswered questions that deserve particular attention. First, is pregnancy-induced central insulin resistance widespread and what are the mechanisms? Here, we might gain clues to the cellular mechanisms based on what is known about peripheral insulin resistance and/or central insulin resistance associated with obesity, although it is also possible these cellular mechanisms may be unique to pregnancy. Furthermore, the factors driving these changes in central insulin sensitivity needs investigation. We know that prolactin and placental lactogens play a key role in changes in insulin secretion during pregnancy ⁸ and hence may also be involved in preventing excessive insulin function in the brain at this time. Indeed, how the hormonal milieu of pregnancy influences central insulin sensitivity will be a key area of investigation in the future. Second, does lactation produce central insulin resistance, and if so, which brain sites and functions are affected? Does lactation alter transport of insulin into the brain or brain insulin levels? Of particular interest will be the investigation of how the transition from pregnancy to lactation impacts on central insulin sensitivity and/or other aspects of central insulin function that may differ between these two states. Finally, we ask, does the same pattern of site-specific central insulin resistance develop in gestational diabetes as in unaffected pregnancies, or perhaps pregnancy-induced central insulin resistance is intensified with this condition as it is in the periphery. Addressing these key questions in the future will expand our knowledge of how the maternal body adapts to these unique reproductive states and allow us to better understand how pregnancy complications may impact on the health and well-being of both mother and baby.