Review pubs.acs.org/molecularpharmaceutics
Intranasal Neuropeptide Administration To Target the Human Brain in Health and Disease Maartje S. Spetter† and Manfred Hallschmid*,†,‡,§ †
Department of Medical Psychology and Behavioral Neurobiology, University of Tübingen, 72076 Tübingen, Germany German Center for Diabetes Research (DZD), 72076 Tübingen, Germany § Institute for Diabetes Research and Metabolic Diseases of the Helmholtz Centre Munich at the University of Tübingen (IDM), 72076 Tübingen, Germany ‡
ABSTRACT: Central nervous system control of metabolic function relies on the input of endocrine messengers from the periphery, including the pancreatic hormone insulin and the adipokine leptin. This concept primarily derives from experiments in animals where substances can be directly applied to the brain. A feasible approach to study the impact of peptidergic messengers on brain function in humans is the intranasal (IN) route of administration, which bypasses the blood−brain barrier and delivers neuropeptides to the brain compartment, but induces considerably less, if any, peripheral uptake than other administration modes. Experimental IN insulin administration has been extensively used to delineate the role of brain insulin signaling in the control of energy homeostasis, but also cognitive function in healthy humans. Clinical pilot studies have found beneficial effects of IN insulin in patients with memory deficits, suggesting that the IN delivery of this and other peptides bears some promise for new, selectively brain-targeted pharmaceutical approaches in the treatment of metabolic and cognitive disorders. More recently, experiments relying on the IN delivery of the hypothalamic hormone oxytocin, which is primarily known for its involvement in psychosocial processes, have provided evidence that oxytocin influences metabolic control in humans. The IN administration of leptin has been successfully tested in animal models but remains to be investigated in the human setting. We briefly summarize the literature on the IN administration of insulin, leptin, and oxytocin, with a particular focus on metabolic effects, and address limitations and perspectives of IN neuropeptide administration. KEYWORDS: insulin, leptin, oxytocin, intranasal administration, central nervous system, brain, metabolism, glucose homeostasis, cognitive function, memory
■
INTRODUCTION The control of energy fluxes by the central nervous system (CNS) strongly depends on the input of peripheral endocrine messengers.1 The pancreatic hormone insulin and the adipocyte-derived hormone leptin circulate at concentrations proportional to body fat stores and, after crossing the blood− brain barrier (BBB) via saturable transport mechanisms,2,3 act as adiposity signals that provide the brain with information on the body’s energy resources. Accordingly, their central nervous administration reduces food intake4,5 and disrupting the brain impact of insulin and leptin induces an obese, diabetic phenotype.6,7 The insulin receptor is expressed in particularly high densities in brain regions like the olfactory bulb, the cerebellum, the dentate gyrus, the pyriform cortex, the hippocampus, the choroid plexus, and the arcuate nucleus of the hypothalamus.8 Peripheral insulin reaching the CNS modulates functions such as energy and glucose homeostasis,5,9 reproduction,6 growth,10 reward processing,11 and neuronal plasticity.12 (Some animal experiments have found indicators for insulin secretion within the CNS,8,13 in particular detecting insulin mRNA in GABAergic neurogliaform cells in the rodent cortex by single-cell digital PCR;14 however, the question of © XXXX American Chemical Society
local insulin production in the human brain clearly is in need of further investigation.) Much like insulin, leptin contributes to processes like the control of glucose homeostasis,15 lipid metabolism,16 and reproductive and cardiovascular function17 as well as reward-related behavior.11 In addition to their roles in the regulation of energy fluxes, leptin as well as insulin and other metabolic endocrine factors such as glucagon-like peptide 1 and ghrelin are also potent modulators of higher cognitive processes including memory formation (for review, see ref 18). The hypothalamic nonapeptide oxytocin, which is released into the circulation by axonal terminals in the posterior pituitary and in addition acts directly on central nervous receptors, has been primarily known for its contribution to reproductive function and mother−infant interaction, and in recent years has gained particular attention as a regulator of Special Issue: Advances in Respiratory and Nasal Drug Delivery Received: January 16, 2015 Revised: April 4, 2015 Accepted: April 16, 2015
A
DOI: 10.1021/acs.molpharmaceut.5b00047 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Review
Molecular Pharmaceutics
weights of 1 kDa, 5.8 kDa, and 16 kDa, respectively, can be expected to differ according to the specific physicochemical properties of the molecules. Passage of IN administered peptides to the brain can also be established along cranial and trigeminal nerve branches,27 and most recently, bulk flow within the perivascular space of cerebral blood vessels has been identified as another transport mechanism following IN administration.32 In this context it is noteworthy that, although the detection of IN peptides in CSF30 was an important step in the translation of the IN paradigm from animals to the human and clinical setting, experiments on the brain uptake of IN leptin33 and insulin-like growth factor-1 (IGF-1)27 suggest that these peptides can effectively reach the brain parenchyma without passage through the CSF compartment. For in-depth information on pharmacological and pharmacodynamic aspects of IN delivery, the reader is referred to refs 29, 34, and 35. Depending on the amount, the concentration, and the physicochemical properties of the specific compound, IN administration can be associated with non-neglible “spillover” permeation into the circulation.35−40 Therefore, recent experiments focusing on brain-mediated effects of IN administered peptides have started to control for potential functional changes in organ function induced by peripheral absorption: see, for example, current studies on the contribution of CNS insulin signaling to glucose homeostasis39,41,42 discussed below. In this context it is important to note, however, that, in accordance with the functional effects of IN neuropeptide delivery outlined in this review, experiments in animals have confirmed that IN administered neuropeptides access brain regions that are crucial for the control of metabolism as well as cognitive function.28,33
psychosocial function in humans (for review, see ref 19). However, experiments in rodents have implicated oxytocin in the control of food intake, energy expenditure, and glucose homeostasis,20,21 functions that according to recent research22,23 also pertain to the human organism. While important insight into the role of insulin, leptin, and oxytocin in the central nervous orchestration of metabolic functions has been obtained in animal experiments, investigating their role in the human brain is hampered by methodological constraints which, however, can be overcome in experimental as well as clinical settings by means of the relatively recent approach of intranasal (IN) neuropeptide delivery. This review briefly introduces the concept of IN neuropeptide delivery and summarizes the effects of IN insulin and oxytocin in healthy humans and patients, along with some remarks on the potential of IN leptin administration. (Note that, as of yet, studies on IN leptin delivery have not been performed in humans; respective animal studies are described below.)
■
THE INTRANASAL PATHWAY In animal experiments, substances can be administered directly to the brain via stereotaxic infusion. This route of delivery cannot be routinely employed in humans. Here, peptides like insulin are usually administered intravenously (IV) to increase their availability in the CNS, as evidenced by an increase in their cerebrospinal fluid (CSF) concentration. The parenteral route implies a number of limitations including but not limited to the induction of peripheral effects that can bias or even counteract the impact on the brain. Thus, both neuronal insulin and leptin receptor activation can affect hormone concentrations in the periphery by, for example, modulating pancreatic insulin secretion.24,25 Also, the strong decrease in blood glucose levels triggered by systemic insulin infusion activates endocrine (stress) axes that can affect brain activity and, below certain thresholds, will strongly impair cognitive function. Insulininduced hypoglycemia can be prevented by the simultaneous infusion of glucose which, however, per se induces effects on CNS activity.18 The euglycemic−hyperinsulinemic clamp procedure to keep blood glucose concentrations stable during IV insulin infusion is time- and labor-intensive and suffers from some methodological drawbacks. For example, IV insulin administration does not permit the differentiation between centrally mediated and direct peripheral effects of insulin on organs like the liver. The methodological obstacles of IV administration can be circumvented using the IN route of administration that has been described in a number of animal paradigms by the group of William Frey26−29 and other researchers. In humans, IN delivered insulin and other peptides like alpha-melanocytestimulating hormone have been shown to reach the CNS within 60 min after administration.30 Considering that the intraneuronal transport of neuropeptides from the nasal mucosa to the olfactory bulb is a process of several hours,26 it is assumed that IN administered neuropeptides reach the CNS largely via extraneuronal passage, bypassing the BBB paracellularly by diffusing into the subarachnoid space across the olfactory epithelia and through intercellular clefts between sustentacular cells and olfactory neurons.29 Peptides can also cross the olfactory epithelium transcellularly, i.e., by passive diffusion (primarily in the case of lipophilic molecules) or receptormediated transcytosis: the olfactory epithelium is, for example, equipped with receptors for insulin.31 Note that nose-to-brain transport of oxytocin, insulin, and leptin, which have molecular
■
INTRANASAL INSULIN MODULATES BRAIN FUNCTION Functional evidence for the efficacy of insulin transport to the human brain after IN administration has been gathered in a series of studies applying different experimental approaches. In early experiments, IN insulin induced distinct alterations in auditory evoked electroencephalographic brain responses recorded in healthy men during an oddball paradigm performed during systemic euglycemia.43 In further studies,44 resting-state brain activity proved to be highly sensitive to IN insulin as indicated by a sustained negative shift in direct current (dc) brain potentials within 20 min after IN insulin administration, which presumably stemmed from the hormone binding to receptors on glial cells. Interestingly, IV insulin administration (irrespective of concurrent eu- or hypoglycemia) yielded comparable effects, suggesting that increases in circulating insulin concentrations are rapidly signaled to the brain. Further experiments relying on functional magnetic resonance imaging (fMRI) have indicated a positive relationship between circulating insulin concentrations and the activation of the right hippocampus when subjects view pictures of high-caloric food items.45 However, in a related fMRI study, food-associated hippocampal activity was attenuated after IN insulin compared to placebo administration.46 In healthy young men, IN insulin administration increased regional cerebral blood flow in insular cortex and putamen,47 which is in line with its role in appetite regulation discussed below. The effect of circulating insulin on cerebrocortical activity under euglycemic conditions was also assessed by means of magnetoencephalographic (MEG) recordings; in those studies, obesity48 and the fat-mass and obesity associated (FTO) allele variant rs8050136 (ref 49) were B
DOI: 10.1021/acs.molpharmaceut.5b00047 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Review
Molecular Pharmaceutics
Figure 1. Postprandial IN insulin administration reduces appetite and snack intake in healthy women. (A) Mean (±SEM) appetite ratings on visual analogue scales anchored at 0 and 100 given by female participants who received IN insulin (160 IU; black dots and solid lines; n = 15) at 1300 h (nose symbol) and control subjects (IN placebo administration; white dots and dashed lines; n = 15). Standard lunch (∼400 kcal) was consumed at 1230 h, and snacks were offered at 1505 h. (B) Mean (±SEM) snack intake (kcal) assessed at 1505 h under the pretext of a taste rating session in the placebo group (white bars) and the insulin group (black bars). Three different types of cookies were offered. (C) Mean (±SEM) snack palatability rated on visual analogue scales anchored at 0 (not palatable) and 100 (highly palatable) during the snack test at 1505 h. * P < 0.05 for comparisons between groups (t tests). Reprinted with permission from ref 53. Copyright 2012 American Diabetes Association.
insulin administration modulates resting-state activity of reward-processing brain circuitries (albeit assessed in the fasted state) in normal-weight women.54 In addition to acting as an adiposity signal and limiting energy intake, insulin in the CNS has also been found to stimulate energy expenditure by increasing sympathetic nervous system outflow to brown adipose tissue57 and inhibiting warmsensitive neurons.58 Accordingly, IN insulin administration enhances postprandial thermogenesis in healthy young men.37 The impact of central nervous insulin on WAT metabolism appears to match the anabolic effect of peripheral insulin in that it inhibits lipolysis and enhances lipogenesis in rodents.59,60 A study in healthy humans corroborated these findings by demonstrating that IN insulin acutely suppresses free fatty acid concentrations and the rate of appearance of deuterated glycerol as an estimate of lipolysis, whereas lipolytic protein expression in subcutaneous adipose tissue remained unchanged.61 In that study, signs of a slight increase in serum insulin concentrations after the IN administration of 160 IU insulin remained nonsignificant and were statistically unrelated to the observed changes in WAT-related parameters, suggesting that the latter were predominantly conveyed via central nervous mechanisms.
found to modulate insulin’s impact on cerebrocortical beta- and theta-wave activity.
■
CNS INSULIN SIGNALING CONTROLS METABOLIC PROCESSES Insulin exerts control of metabolic functions by binding to insulin receptors on peripheral organs, e.g., limiting postprandial blood glucose excursions by stimulating glucose uptake into muscle and white adipose tissue (WAT) as well as glycogen synthesis while reducing hepatic gluconeogenesis. Moreover, the hormone triggers de novo lipogenesis in WAT and liver and inhibits hepatic triglyceride secretion, thereby boosting fat storage. Studies in animals have indicated that many of these peripheral functions of the hormone are strongly influenced by CNS insulin signaling in a functionally redundant or, in some instances, opposing fashion (for review, see ref 50). Thus, whereas systemic insulin acts as an anabolic hormone by promoting body weight gain in the form of muscle and fat mass, the CNS administration of insulin induces catabolic effects by reducing food intake and, in the long run, body fat content.4,5 In humans, behavioral51−53 as well as neuroimaging studies46,54 relying on the IN paradigm have corroborated the concept of insulin as an important anorexigenic and catabolic signal in the CNS control of ingestive behavior and highlighted potential clinical implications. Healthy humans eat less52,53 and reduce their body weight and body fat content51 when receiving, respectively, acute and eight-week IN insulin treatment. These effects have been observed to show a preponderance in male in comparison to female subjects,51,52 which is in accordance with results in rats55 and might be due to sex differences in the storage of body fat. Although animal data have indicated that estrogen signaling might modulate the sensitivity of the CNS to insulin’s anorexigenic impact,56 studies in humans failed to indicate respective differences in central nervous insulin sensitivity between postmenopausal and young women.36 Interestingly, when administered postprandially, IN insulin intensifies satiety and curbs the consumption and rated palatability of snacks in healthy women (Figure 1),53 suggesting that meal-related insulin secretion acts as a satiety signal and might in particular affect the reward-related component of eating. This conclusion is in accordance with findings that IN
■
DOES INSULIN SIGNALING IN THE CNS REGULATE PERIPHERAL GLUCOSE FLUXES IN HUMANS? Hepatic glucose metabolism is a key element in the maintenance of euglycemia. The liver keeps plasma glucose concentrations stable under conditions of (postprandial) glucose abundance and (fasting) glucose depletion by, respectively, glycogenesis and glycogenolysis/gluconeogenesis. These processes have been traditionally assumed to be governed by direct insulin action on hepatic insulin receptors and indirect insulin effects on liver functions, such as the downregulation of glucagon secretion and plasma nonesterified fatty acid levels (for review see ref 62). Recent evidence suggests that an insulin-dependent brain−liver axis might also play a role in this glucoregulatory network. In rodents, the genetic downregulation of hypothalamic insulin receptors increases hepatic glucose production,63 indicating that the inhibition of hypothalamic insulin signaling decreases hepatic C
DOI: 10.1021/acs.molpharmaceut.5b00047 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Review
Molecular Pharmaceutics
arterial pancreatic clamp (which prevents fluctuations in insulin and glucagon concentrations) using a primed, constant infusion of deuterated glucose was observed to suppress endogenous glucose production by 35% around 180 min after delivery. Venous insulin concentrations were kept constant across conditions by means of IV infusion of a small dose of insulin lispro in the placebo condition, and the authors tentatively conclude that IN insulin downregulates hepatic glucose production via central nervous pathways, although a contribution of a reduction in circulating free fatty acid concentrations emerging after IN insulin delivery cannot be excluded. Related studies in humans have even suggested that this presumed brain−pancreatic cross-talk might be accessible to classical conditioning.70 Clamp studies are the standard experimental approach of measuring insulin sensitivity but imply a decrease in the physiological gradient between portal vein and systemic/ cerebral insulin exposure, resulting in relative hepatic insulin deficiency.71 Further recent experiments have therefore investigated whether acute increases in brain insulin signaling impact peripheral glucose homeostasis under normal fasting conditions.41,42 In the study by Ott and colleagues,41 20 fasted healthy men IN received up to 420 IU of the insulin analogue aspart during experimental sessions lasting 6 h. IN insulin aspart dose-dependently decreased plasma glucose concentrations, but when the concurrent leakage of the IN administered compound into the circulation was mimicked by IV infusion in control experiments, essentially the same reduction of blood glucose concentrations was observed. In related experiments, IN insulin administration did not affect endogenous glucose production and hepatic insulin sensitivity in fasted healthy subjects and patients with type 2 diabetes.42 In line with findings in dogs,64,72 these results argue against a pivotal contribution of brain insulin signaling to the acute control of glucose fluxes under fasting conditions in humans. Still, it cannot be excluded that brain insulin effects might develop across longer time periods, although the observation that the regulation of hepatic glucose production is maintained in liver-transplanted patients 73 does not support this assumption. Clearly, there is a need for additional research on the role of CNS insulin signaling in human glucose metabolism and its role in the development of diabetes.74 In these studies, it will be of great importance to rely on appropriate methodology including adequate control conditions.
insulin sensitivity. Fittingly, insulin hyperpolarizes glucoseresponsive hypothalamic neurons by opening ATP-sensitive potassium (KATP) channels, triggering the vagal transmission of a signal that induces the suppression of hepatic glucose production.9,63 However, these findings in rodents have been challenged by experiments in dogs where a 4-fold increase in the concentration of circulating insulin selectively in brainafferent arteries did not potentiate the inhibition of hepatic glucose production.64 In accordance, intracerebroventricular (ICV) insulin administration to dogs, although reducing hepatic gluconeogenic gene expression, did not alter hepatic glucose production.65 Interestingly, however, canine experiments performed over 30 years ago showed that ICV insulin administration increases pancreatic insulin secretion via a feedforward mechanism, pointing to the relevance of vagal brain− pancreas crosstalk also in higher organisms.25,66 Supporting this assumption, IV insulin administration in humans has been observed to trigger a positive feedback mechanism on glucoseinduced pancreatic insulin secretion.67 The question whether brain insulin signaling might be a significant regulator of peripheral glucose metabolism in humans has been addressed in a series of recent studies. Orally administering the KATP channel opener diazoxide to healthy subjects was observed to decrease hepatic glucose production assessed by tracer dilution techniques during a basal insulin clamp.68 In control experiments in rats that showed basically the same response, ICV infusion of the KATP channel blocker glibenclamide abolished the diazoxide-induced reduction in hepatic glucose production, buttressing the conclusion that the observed effects were centrally mediated. Although systemic effects of diazoxide, which potently inhibits insulin secretion via peripheral effects, cannot be ruled out, the results of this study argued for the presence of a brain−liver axis also in humans. Accordingly, IN insulin administration to healthy men before meal intake decreased the postprandial surge in circulating insulin concentrations while plasma glucose excursions corresponded to the level observed in the placebo condition,37 pointing toward a beneficial effect of brain insulin administration on peripheral insulin sensitivity. This interpretation is supported by analyses of fasting plasma glucose and insulin concentrations in over 100 participants which suggest that IN insulin improves systemic insulin sensitivity.38 This hypothesis was directly tested in further experiments by Heni and colleagues69 where peripheral insulin sensitivity was assessed by means of a hyperinsulinemic euglycemic glucose clamp 90 min before until 120 min after the IN administration of 160 IU insulin to ten normal-weight and five obese subjects. In lean but not in obese participants, the glucose infusion rate to maintain euglycemia increased within around 60−90 min after IN insulin compared to placebo administration, suggesting an increase in whole-body insulin sensitivity. This effect was positively related to IN insulin-induced changes in heart rate variability and fMRI-assessed hypothalamic blood flow and, therefore, assumed to have been mediated via hypothalamic−parasympathetic downstream signaling pathways. In accordance with previous studies on IN insulin administration,36−38 a slight and transient increase in circulating insulin concentrations emerged within 15 min after IN insulin delivery, most probably reflecting spillover of IN delivered insulin into the bloodstream. The possible contribution to changes in systemic insulin sensitivity of IN insulin reaching the circulation was controlled for in experiments by Dash and co-workers.39 Here, 40 IU of IN insulin lispro administered to eight healthy men during an
■
LEPTIN AND OXYTOCIN IN THE CONTROL OF METABOLIC FUNCTION The contribution of leptin to the control of metabolic function, namely, its role as an adiposity signal to the brain, and its interplay with other metabolic messengers have been extensively investigated.75 Compared to insulin, CNS leptin signaling exerts opposing effects on glucose and lipid metabolism, promoting gluconeogenesis and stimulating lipolysis, but blocking de novo lipogenesis in WAT, i.e., supporting catabolic processes. These metabolic effects of leptin appear to be primarily conveyed via central nervous pathways, which is illustrated by the fact that an induced knockout of the leptin receptor exclusively in the body periphery does not entail an obese or diabetic phenotype in mice.76 In contrast, the targeted reconstitution of leptin receptors in the CNS reverses the obese, hyperphagic phenotype of leptin receptor-deficient db/db mice.77 Evidence D
DOI: 10.1021/acs.molpharmaceut.5b00047 Mol. Pharmaceutics XXXX, XXX, XXX−XXX
Review
Molecular Pharmaceutics
Figure 2. IN oxytocin inhibits reward-driven eating in healthy men. (A) Mean (±SEM) hunger ratings assessed before and after IN administration (upright dotted line) of oxytocin (24 IU; black dots and solid lines) and placebo (vehicle; white dots and dotted lines). 45 min post-treatment, fasted subjects ate from a test breakfast (1000−1030 h; total food intake, placebo condition, 1180 ± 103 kcal; oxytocin condition, 1190 ± 105 kcal; P > 0.84), and 100 min thereafter, they ingested snacks under the pretext of a taste test (1240−1250 h). (B) Mean (±SEM) cumulative snack intake in the placebo (white bar) and the oxytocin condition (black bar). (C) Individual chocolate cookie consumption assessed at the same test in the placebo (left) and the oxytocin condition (right). Individual values of both sessions are connected by lines. n = 20; *P < 0.05, **P < 0.01 for comparisons between conditions (paired t tests). Reprinted with permission from ref 22. Copyright 2013 American Diabetes Association.
This assumption is supported by our experiments in humans.22 Here, the IN administration of oxytocin to normal-weight, healthy fasted men did not alter hunger-driven food intake from a large breakfast buffet, which is in accordance with findings in animals,91 but strongly decreased the consumption of chocolate cookies taking place around 1.5 h later, when reward-related eating motivation prevailed (Figure 2). Moreover, in accordance with experiments in humans92 and animals,20,93 respectively, IN oxytocin suppressed endocrine stress axis activity and curbed the postprandial peak in plasma glucose concentrations. It remains to be tested if the effect of oxytocin on reward-driven eating can be attributed to a more general impact on reward-related behaviors.94,95 Still, these findings open up an interesting new perspective for oxytocin as a regulator of eating behavior in humans, which might be interrelated with its psychosocial function particularly in social settings of food intake. In this context, it is interesting to note that patients with Prader−Willi syndrome who suffer from hyperphagic obesity because of insatiable food craving have been found to display a 40% reduction in the number and size of oxytocin neurons.96 First pilot trials on oxytocin substitution via the IN pathway in these patients (spanning an intervention period of 8 weeks) yielded none of the intended effects on body weight and psychosocial function, which might have been due to lack of feed-forward endogenous oxytocin release after exogenous delivery.97
for a critical role of brain leptin signaling in metabolic control and, in particular, the regulation of eating behavior also derives from studies in humans. In obese compared to lean adolescents, brain responses to pictures of high-calorie foods were found to be increased in reward-processing striatal−limbic structures and, notably, were positively correlated with endogenous leptin concentrations.78 Accordingly, leptin-deficient humans who reduce their body weight during leptin replacement therapy show increased neuronal activation in the prefrontal cortex,79 which establishes inhibitory control of eating behavior, and changes in the activity of striatal regions that regulate rewardrelated behavior.80 Also, normal-weight women with acquired hypoleptinemia (defined as fasting leptin levels
