Satya Dash, Changting Xiao, Cecilia Morgantini, Khajag Koulajian, and Gary F. Lewis

Intranasal Insulin SuppressesEndogenous GlucoseProduction in HumansCompared With Placebo in thePresence of Similar VenousInsulin ConcentrationsDiabetes 2015;64:766–774 | DOI: 10.2337/db14-0685

Intranasal insulin (INI) has been shown to modulate foodintake and food-related activity in the central nervoussystem in humans. Because INI increases insulin concen-tration in the cerebrospinal fluid, these effects have beenpostulated to be mediated via insulin action in the brain,although peripheral effects of insulin cannot be excluded. INIhas been shown to lower plasma glucose in some studies,but whether it regulates endogenous glucose production(EGP) is not known. To assess the role of INI in the regulationof EGP, eight healthy men were studied in a single-blind,crossover study with two randomized visits (one with 40 IUINI and the other with intranasal placebo [INP] administra-tion) 4 weeks apart. EGP was assessed under conditions ofan arterial pancreatic clamp, with a primed, constant in-fusion of deuterated glucose and infusion of 20% dextroseas required to maintain euglycemia. Between 180 and 360min after administration, INI significantly suppressed EGP by35.6% compared with INP, despite similar venous insulinconcentrations. In conclusion, INI lowers EGP in humanscompared with INP, despite similar venous insulin concen-trations. INI may therefore be of value in treating excess liverglucose production in diabetes.

Dysregulation of insulin-mediated suppression of hepaticglucose production (HGP) is a hallmark of type 2 diabetes

(1). It is well established that activation of hepatic insulinreceptors with ensuing activation of downstream insulinsignaling pathways lowers hepatic glucose output by decreas-ing gluconeogenesis and increasing glycogen synthesis (1). Inaddition to the direct action of insulin on hepatocytes, in-sulin can indirectly affect HGP by altering free fatty acid(FFA) flux and suppressing glucagon secretion (2–4). Animalstudies have indicated that insulin may also indirectly regu-late hepatic glucose output via effects on the central ner-vous system (CNS)—the so-called brain–liver axis (1,5,6).

Injection of insulin into the third cerebral ventriclein mice activates KATP channels in the mediobasalhypothalamus (via activation of the insulin receptor–phosphoinositide 3-kinase pathway), which activatessecond-order neurons in the brain stem. This in turnlowers expression of gluconeogenic enzymes and glucoseproduction by the liver, an effect that is abrogated bysurgical resection of the hepatic branch of the vagusnerve (6). CNS insulin action in the dorsal vagal complexhas more recently been shown to regulate HGP via activa-tion of the insulin–insulin receptor–extracellular signal–related kinase pathway (7). These brain–liver axis effectsare observed in the absence of changes in plasma insulinconcentration (5–7). A brain–liver axis also has been

Departments of Medicine and Physiology and the Banting and Best DiabetesCentre, University of Toronto, Toronto, Ontario, Canada, and Division of Endo-crinology, Department of Medicine, University of Toronto, Toronto, Ontario,Canada

Corresponding author: Gary F. Lewis, gary.lewis@uhn.ca.

Received 30 April 2014 and accepted 2 October 2014.

Clinical trial reg. no. NCT02131948, clinicaltrials.gov.

This article contains Supplementary Data online at http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1.

S.D. and C.X. contributed equally to this work.

© 2015 by the American Diabetes Association. Readers may use this article aslong as the work is properly cited, the use is educational and not for profit, andthe work is not altered.

See accompanying article, p. 696.

766 Diabetes Volume 64, March 2015

METABOLISM

http://crossmark.crossref.org/dialog/?doi=10.2337/db14-0685&domain=pdf&date_stamp=2015-02-11mailto:gary.lewis@uhn.cahttp://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1

demonstrated in dogs. Insulin delivery into the headarteries augments hepatic glycogen synthesis and reducesmRNA expression of gluconeogenic enzymes but causes noacute change in HGP (8).

In recent human studies a single dose of intranasalinsulin (INI; 160 IU) modulated food-related activity inthe CNS (9), reduced overall food intake in males (10),improved cognitive performance in females (10), and re-duced intake of palatable food and increased satiety infemales (11). These effects are likely mediated by directaction of insulin on CNS, although other non-CNS effectsof nasally administered insulin cannot be definitively ex-cluded. At a lower dose of 40 IU, INI increased the cere-brospinal fluid insulin concentration, with no significantchange in measured serum insulin concentration sampledat 10-min intervals for the first 40 min and at 20-minintervals for a further 40 min (12). Small lipophilic pep-tides such as insulin can potentially enter the CNS directlyby diffusing across the olfactory epithelia and intercellularspaces into the subarachnoid space (12,13). Intranasallyadministered peptides can also enter the CNS indirectlyvia uptake into the olfactory bulb and axonal transport(12,13). Recent studies in humans showed that INI ad-ministered at a higher dose of 160 IU can acutely lowerplasma glucose and alter peripheral insulin sensitivity(11,14,15), an effect postulated to occur via insulin deliv-ery to the CNS. However, the higher dose of INI (160 IU)transiently increased peripheral insulin concentration

(9,14,15), which may contribute to the acute effects ofINI on peripheral glucose metabolism and insulin sensi-tivity (11,14,15).

In this single-blind, placebo-controlled, crossover study,we aimed to investigate whether INI action regulates endog-enous glucose production (EGP) in humans. We assessedEGP following the administration of 40 IU of INI orintranasal placebo (INP) with primed, constant infusion ofd-[6,69-2H2]glucose (D2-glucose) (Fig. 1). Participants werestudied under conditions of an arterial pancreatic clamp inwhich systemic venous insulin and glucagon concentrationsare clamped at basal concentrations to prevent fluctuationsin peripheral arterial insulin and glucagon concentrations.Lower portal insulin and glucagon concentrations thanwould be expected in the basal state—that is, this is a stateof hepatic insulin and glucagon deficiency—are noted withthis clamp technique. Under these conditions we demon-strated for the first time in humans that INI lowers EGP.This effect was seen in the presence of similar venous in-sulin concentrations.

RESEARCH DESIGN AND METHODS

Study ParticipantsEight healthy men with no medical illnesses and taking nomedications were recruited by advertisements in the localpress. Their demographic and biochemical parameters areshown in Table 1. They underwent a 75-g oral glucosetolerance test, routine screening blood tests, and urinalysis.

Figure 1—Outline of the study. Participants were admitted to the Metabolic Test Centre the day before the study. They were given a mixedmeal at 5 P.M. At 7 A.M. (t = 2120 min) a primed, constant infusion of D2-glucose was started and continued for the duration of the study. Atthe same time, an arterial pancreatic clamp, with infusion of somatostatin along with replacement doses of insulin, glucagon, and growthhormone, was started as described in the RESEARCH DESIGN AND METHODS. At 9 A.M. (t = 0 min) 40 IU of INI or placebo was administered. †Toensure similar venous insulin concentration between treatments, the study subjects receiving placebo were given 0.005 IU/kg insulin lisprointravenously (i.v.) over 30 min, starting at the same time as the intranasal placebo because there was spillover of INI. NPO, nil per os(nothing by mouth).

diabetes.diabetesjournals.org Dash and Associates 767

Those with abnormal tests were excluded. Each participantwas studied on two occasions, 4 weeks apart, in a single-blind, placebo-controlled, crossover trial. Each participantreceived 40 IU of INI lispro (Humalog; Eli Lilly Canada,Toronto, Ontario, Canada) during one visit and placeboduring the other. The order of the visits was randomized.

Insulin DosingPilot studies indicated that peripheral insulin spillover wasinevitable, with doses ranging from 80 IU to as low as 10IU (Supplementary Fig. 1). A dose of 40 IU INI was chosenbecause in previous studies it increased cerebrospinal fluid(CSF) insulin concentration in humans (12), and in ourpilot studies a higher dose of INI (80 IU) resulted ina greater spillover of insulin lispro into the peripheral cir-culation (Supplementary Fig. 1). To ensure similar venousplasma insulin concentrations between treatments, 0.005IU/kg insulin lispro was infused intravenously over 30 min,starting at the time of INP administration. The intravenousinsulin lispro dose was identified in a pilot study usingvaried doses (data not shown).

Study OutlineParticipants were admitted to the Metabolic Test Centreof Toronto General Hospital the morning before thestudy (Fig. 1). Volunteers had a standardized mixed mealat 5 P.M. and were not permitted any food or drink orallyexcept water until the conclusion of the study. The fol-lowing day at 7 A.M. (t = 2120 min), an arterial pancreaticclamp was started and continued for the remainder ofthe study (until 3 P.M.; t = 360 min) to neutralize anypotential effects of arterial pancreatic hormone fluctu-ations on EGP. The clamp comprised the following in-fusions: somatostatin 30 mg/h (Sandostatin; NovartisPharmaceuticals Canada, Dorval, Quebec, Canada) toinhibit pancreatic insulin and glucagon secretion withconcomitant replacement at basal concentrations of in-sulin (Humulin R; Eli Lilly Canada) at 0.05 mU/kg/min;human recombinant growth hormone (Humatrope; EliLilly Canada) at 3 ng/kg/min; and glucagon (Eli LillyCanada) at 0.325 ng/kg/min. Autologous serum (3 mL),freshly prepared from the subject’s blood, was added tothe saline as a carrier before hormone dilution.

Also at 7 A.M. (t = 2120 min), a primed, constant infu-sion (22.5 mmol/kg bolus followed by 0.25 mmol/kg/min)of D2-glucose (Cambridge Isotope Laboratories, Tewksbury,MA) was started and continued until the conclusion of

the study at 3 P.M. (t = 360 min). At 9 A.M. (t = 0 min), par-ticipants received either 40 IU of insulin lispro (100 IU/mLHumalog; Eli Lilly Canada) or placebo (insulin diluent; EliLilly Canada) via a metered nasal dispenser (Pharmasystems,Markham, Ontario, Canada). A single spray dispenses 0.1mL (10 IU) of insulin. A single spray was administered ineach nostril while inhaling. Another spray was similarlyadministered in each nostril 1 min later to give a totaldose of 40 IU of insulin lispro. Because of spillover of 40IU insulin lispro into the peripheral circulation (Supple-mentary Fig. 1), volunteers received an intravenous in-fusion of insulin lispro during the placebo visit to ensuresimilar venous insulin concentration in both treatmentgroups, as described above.

Blood samples (10 mL) were collected every 30 min forthe first 120 min (t = 2120 to 0 min) after starting thearterial pancreatic clamp, every 5 min for 30 min afteradministration of INI/INP (t = 0 to 30 min), and every 10min thereafter until the conclusion of the study (t = 30 to360 min). A 20% dextrose solution was administered asnecessary to maintain euglycemia.

Laboratory MethodsPlasma was separated from blood samples in a refrigeratedcentrifuge at 3,000 rpm for 15 min at 4°C. Sodium azide(70 mg/L blood; Sigma-Aldrich, Oakville, Canada) andaprotinin (1.94 mg/L blood; Sigma-Aldrich) were addedto the plasma to prevent hydrolysis and protein degrada-tion. Plasma was dried and derivatized, and stable isotopeenrichments were determined (16). Derivatized sampleswere analyzed with gas chromatography/mass spectrom-etry (Agilent 5975/6890N; Agilent Technologies CanadaInc., Mississauga, Ontario, Canada) with electron impactionization using helium as the carrier gas. Selective ionmonitoring with a charge-to-mass ratio of 242 and 244was performed. Atom percentage of excess fraction (APE)was calculated for each sample as APE = tracer/(tracer +tracee).

Commercial kits were used to measure total insulin(Millipore, Billerica, MA), growth hormone (Abcam Inc.,Toronto, Ontario, Canada), FFA (Wako Industrials, Osaka,Japan), triglyceride (TG) (Roche Diagnostics), and gluca-gon (Millipore). An insulin lispro kit (Millipore) (specific-ity for lispro 100%, specificity for human insulin;0.05%)was used to measure lispro concentration in our pilotstudies (Supplementary Fig. 1).

Analysis of EGPEGP was calculated as described previously (17). Duringa steady state, the rate of glucose appearance (Ra) wasequivalent to the rate of glucose disappearance (Rd),where Ra = tracer infusion rate/APE fraction and EGPrate = Ra – glucose infusion rate (17).

StatisticsResults are presented as mean 6 SEM. Paired t tests wereused to compare plasma glucose, venous insulin, plasmaglucagon, glucose infusion rates, and EGP. A P value

Table 1—Baseline demographics and biochemistry

Characteristics Mean 6 SEM

Age (years) 49.1 6 2.0

Body weight (kg) 73.9 6 2.7

BMI (kg/m2) 23.9 6 0.8

Fasting plasma glucose (mmol/L) 4.9 6 0.1

Fasting plasma insulin (pmol/L) 40.6 6 6.0

768 INI and Endogenous Glucose Production Diabetes Volume 64, March 2015

http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1

,0.05 was considered significant. Post hoc analysisrevealed a power of 98% to detect a change in EGP aswell as plasma insulin in the final 180 min of the study(t = 180–360 min). The power to detect a change in in-sulin concentration during the entire study was 94%.

Study OversightThe study was carried out according to the principles ofthe Declaration of Helsinki and was approved by Univer-sity Health Network Research Ethics Board, Toronto,Ontario, Canada. All participants gave written informedconsent.

RESULTS

Plasma Glucose and Insulin ConcentrationsMean glucose concentration over time during the study isdepicted in Fig. 2A. INI treatment transiently loweredplasma glucose concentrations, with a mean nadir concen-tration at 180 min (INP 6.1 6 0.3 vs. INI 5.1 6 0.1mmol/L; P = 0.01) (Fig. 2A). With infusion of 20%

dextrose, blood glucose concentrations were not signifi-cantly different in the final 120 min of the study. Therewas no significant difference in mean venous plasma in-sulin concentration (Fig. 2B) between treatments (INP71 6 7 vs. nasal insulin 77 6 11 pmol/L; P = 0.6).As expected, based on pilot data, there was a transientincrease in venous insulin after INI administration(Supplementary Fig. 1). Administration of a 30-min in-travenous insulin lispro infusion along with INP admin-istration ensured venous insulin concentrations weresimilar between the groups.

INI Treatment Increases Intravenous Glucose InfusionRequirements to Maintain EuglycemiaIntravenous glucose in the form of 20% dextrose wasinfused, as required, to maintain euglycemia. The meandextrose infusion rate over time is illustrated in Fig. 2C.The mean dextrose infusion rate in the final 180 minof the study (t = 180–360 min) was significantly higher

Figure 2—A: Mean plasma glucose concentrations over time during the course of the study (INP: –◇–; INI: –■–). INI treatment loweredplasma glucose concentrations, with a mean nadir concentration at 180 min. With infusion of 20% dextrose to maintain euglycemia (asshown in C), blood glucose concentrations were not significantly different between treatments in the final 120 min of the study. B: Meanvenous insulin concentrations over time during the course of the study. (INP: –◇–; INI: –■–). There were no significant differences invenous insulin concentration between treatments. Consistent with our pilot data (Supplementary Fig. 1), there was a transient increasein venous insulin concentration with INI, which was mimicked in the INP group by administration of intravenous insulin lispro from 0 to30 min. C: Time course of mean glucose infusion (20% dextrose in mmol/min/kg) required to maintain euglycemia during the study (INP:–◇–; INI: –■–). Glucose infusion rates increased after ;160–180 min with INI treatment, with the maximal difference in infusion rate at 250min (P = 0.006). D: Mean glucose infusion rates over the final 180 min (t = 180–360 min) of the study (INP: white bar; INI: black bar). Meanglucose infusion rates were significantly higher with INI treatment (P = 0.015).

diabetes.diabetesjournals.org Dash and Associates 769

http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1

with INI treatment (placebo 1.1 6 0.9 vs. INI 6.6 6 1.6mmol/min/kg; P = 0.015) (Fig. 2D); the maximal differ-ence occurred at 250 min (placebo 1 6 1 vs. INI 8.4 61.1 mmol/min/kg; P = 0.006) (Fig. 2C). These changesoccurred despite similar venous insulin concentrations.

INI Suppresses EGP Without AffectingGlucose DisposalMean EGP rate over time is shown in Fig. 3A. INI loweredEGP, with the nadir value at 250 min (placebo 12.3 6 1.3vs. INI 6 6 0.9 mmol/min/kg; P = 0.01). Mean EGP in thefinal 180 min of the study (t = 180–360 min) was signif-icantly lower with INI treatment (placebo 11.7 6 1 vs. INI7.6 6 0.6 mmol/min/kg; P = 0.02) (Fig. 3B). The correla-tion coefficient between peak insulin after INI or placeboadministration and decline in EGP from baseline in thefinal 180 min of the study was 0.38.

The rate of glucose disposal (Rd) over time is depictedin Fig. 3C. There was no significant difference in glucosedisposal. Mean Rd in the final 180 min of the study wasnot different between the groups (placebo 12.9 6 1.2 vs.INI 13.76 1 mmol/min/kg; P = 0.4) (Fig. 3D). The specific

activity of D2-glucose over time is shown in Supplemen-tary Fig. 3.

Plasma FFA and TG ConcentrationPlasma FFA concentrations are shown in Fig. 4A. FFAconcentration was significantly lower at 240 min withINI (placebo 0.2 6 0.05 vs. INI 0.1 6 0.03 mmol/L; P =0.03). Mean FFA concentration in the final 180 min (t =180–360 min) was not significantly different betweentreatments (placebo 0.20 6 0.05 vs. INI 0.13 6 0.03mmol/L; P = 0.12) (Fig. 4B).

Plasma TG concentrations are shown in Fig. 4C. Therewas no significant difference in mean TG concentrationfrom t = 180 to t = 360 min (placebo 0.7 6 0.1 vs. INI0.6 6 0.1 mmol/L; P = 0.23) (Fig. 4D).

DISCUSSION

Rodent studies have demonstrated that insulin action inthe CNS can reduce HGP (5,6). Previous human studiesthat deployed INI at a dose that increases CSF insulinconcentration (12) reported changes in peripheral glu-cose concentration and insulin sensitivity, suggesting

Figure 3—A: Time course of mean rates of EGP during the study (INP: –◇–; INI: –■–). EGP rates declined from ;180 min with INItreatment, with a nadir at 250 min (P = 0.01). B: Mean EGP in the final 180 min of the study (t = 180–360 min) (INP: white bar; INI: black bar).Mean EGP was lower with INI treatment (P = 0.02). C: Time course of mean Rd rates during the study (INP: –◇–; INI: –■–). No significantdifference between treatments was observed. D: Mean Rd in the final 180 min of the study (t = 180–360 min) (INP: white bar; INI: black bar).No significant difference between treatments was observed.

770 INI and Endogenous Glucose Production Diabetes Volume 64, March 2015

http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1

CNS insulin may regulate peripheral glucose metabolism(11,14,15). In these studies, however, there was a tran-sient increase in venous insulin concentration after INIadministration. The current study is the first humanstudy to definitively demonstrate that INI (40 IU) sup-presses EGP compared with INP. Importantly, the exper-imental design of our study ensured that venous insulinconcentrations were similar between treatments.

A 40-IU dose of INI previously caused a rapid increasein CSF insulin concentration (12) without increasingserum insulin. In this study, under conditions of an ar-terial pancreatic clamp (during which endogenous insu-lin secretion cannot be modulated), 40 IU of INI (insulinlispro; Eli Lilly Canada) also transiently increased venousinsulin concentration, as measured by a specific lisproassay. We also detected peripheral spillover with a doseas low as 10 IU (Supplementary Fig. 1). In this study weadministered the lowest dose of nasal insulin that waspreviously shown to increase CSF insulin concentration

(12) while minimizing systemic spillover by not usinga higher dose. Subjects treated with INP were given aninfusion of insulin lispro to try and mimic the increasein venous insulin concentrations after spillover of INI(Fig. 1). This experimental approach ensured similar ve-nous insulin concentrations between the treatment groups.

The peripheral spillover of INI, detected with frequentblood sampling (every 5–10 min), may have implicationsin the interpretation of certain findings from previousstudies using higher doses of INI. Three studies usedhigher doses of INI (160 IU) and reported relatively mod-est lowering of plasma glucose concentration within 30–45 min of administration (11,14), as well as improvedperipheral insulin sensitivity (15). These studies did notmeasure plasma insulin as frequently as we did in thecurrent study. In one study (14), plasma insulin (mea-sured every 30 min) was significantly higher at 30 min,with a decline in C-peptide and subsequent insulin concen-tration, suggesting peripheral spillover of insulin with

Figure 4—A: Time course of plasma FFA concentration during the study (INP: –◇–; INI: –■–). FFA concentration was significantly lowerwith INI at 240 min (P = 0.03). B: Mean plasma FFA concentration over the final 180 min (t = 180–360 min) of the study (INP: white bar; INI:black bar). No significant difference between treatments was observed. C: Time course of plasma TG concentration during the study (INP:–◇–; INI: –■–). No significant difference between treatments was observed. D: Mean plasma TG concentration over the final 180 min (t =180–360 min) of the study (INP: white bar; INI: black bar). No significant difference between treatments was observed.

diabetes.diabetesjournals.org Dash and Associates 771

http://diabetes.diabetesjournals.org/lookup/suppl/doi:10.2337/db14-0685/-/DC1

a decline in endogenous insulin secretion. Plasma insulinconcentration (measured every 15 min) was transientlyhigher at 15 min after administration of 160 IU of INIin a recent study by the same group (15). Insulin concen-tration was not reported for the first 45 min in the studyby Hallschmid et al. (11). In these studies of INI, excludinga contribution of peripheral insulin action to the rapid de-crease in plasma glucose concentration is not possible. It isworth noting that other studies using 160 IU of INI did notreport changes in plasma glucose concentrations (9,18).This may be because of the relative infrequency of bloodsampling (every 30 min compared with every 5–10 min inthe current study) (11). In addition, in the absence of anarterial pancreatic clamp, endogenous insulin and glucagonsecretion can be modulated to prevent major fluctuationsin plasma glucose (9,14). With an arterial pancreatic clamp,the normal portal to peripheral insulin gradient is lost,resulting in relative hepatic insulin deficiency, whichmay have permitted INI to lower EGP. In the aforemen-tioned studies (9,14) the physiological portal peripheralgradients of insulin and glucagon were maintained; thismay have rendered the liver less sensitive to INI. Finally,these studies were of a shorter duration (#180 min)(9,11,14,18) than the current study (360 min), in whichEGP declined after 180 min, and therefore a late glucose-lowering effect would have gone undetected.

Unlike the relatively rapid decrease of plasma glucoseseen with INI in some studies (11,14,15), extra-pancreaticKATP channel activation (likely CNS KATP channel activa-tion, a downstream target of insulin action in the CNS)previously decreased EGP in humans over the course of;6–7 h; parallel studies using rats demonstrated that anequivalent dose of diazoxide is detectable in CSF after 1 hand plateaus at ;4 h (16). We therefore speculate thatthe rapid decrease in plasma glucose that occurred with160 IU in previously published studies (11,14) was likelydue to transient systemic insulin absorption from the INIadministration, which in turn induces peripheral insulinaction, and that any potential CNS regulation of EGP byinsulin would be a slower process. Consistent with ourhypothesis, despite similar venous insulin concentrationsin both treatment groups throughout the study, INI low-ered EGP after ;180 min, and EGP remained lower atthe conclusion of the study (at 360 min). This time scaleof INI action is similar to that reported with intracere-broventricular injection of insulin in rodents (5,6) andsuggests that INI, potentially via CNS insulin action,does not rapidly regulate EGP in the acute setting. It isnot known whether prior exposure to INI, as occurs withlonger-term administration of INI, affects EGP, but 8weeks of treatment with INI did not affect fasting in-sulin and glucose (19).

Although the exact mechanism by which INI lowersEGP remains to be determined, reduced expression ofhepatic gluconeogenic enzymes such as PEPCK andG6Pase via hepatic vagal efferents secondary to CNSinsulin action is a plausible explanation. This is based on

the previous findings that 1) INI increases CSF insulinconcentration (12) and 2) intracerebroventricular insulininjection reduces the expression of gluconeogenicenzymes and HGP in rodents, an effect abrogated byresection of hepatic vagal efferents (5,6). A previousstudy demonstrated a reduction in FFA with 160 IU ofINI treatment in humans with no reported change inplasma insulin concentration (18), although as dis-cussed above at this dose peripheral insulin spillovermay have contributed. In the current study, mean FFAbetween 180 and 360 min tended to be lower with INIbut did not reach statistical significance. In view of therelatively small sample size of the current study, wecannot exclude a significant difference in FFA had westudied more individuals, and therefore reduced FFAflux to the liver could potentially contribute to the low-ering of HGP. There were no significant differences in theconcentration of growth hormone and glucagon betweentreatments. Previous studies of dogs demonstrated thatpulmonary insulin delivery can regulate insulin sensitiv-ity independent of plasma insulin concentration (20).Nasally inhaled aerosols are deposited in the lungs (21).INI can potentially deliver insulin to the lungs with sec-ondary effects on glucose metabolism. It is not possibleto rule out as yet unidentified CNS- and/or non-CNS-mediated effects as contributors to EGP reduction.

Although this study suggests that INI, possibly actingvia the CNS, can regulate EGP, the contribution of thispathway to glucose homeostasis in normal human phys-iology remains to be determined. Under our arterialpancreatic clamp conditions, the pancreas is unable tomodulate endogenous insulin and glucagon secretion withchanging glucose concentration. In addition, with anarterial pancreatic clamp, as was the case in previousrodent studies (5,6), portal and peripheral insulin con-centrations are likely to be identical because the normalportal-peripheral insulin concentration gradient is abol-ished when insulin delivery does not occur via the portalcirculation. In normal physiology, portal insulin concen-tration is nearly three times greater than its peripheralconcentration (8,22). Hence during an arterial pancreaticclamp there is relative hepatic insulin deficiency (8,22).Under these conditions, CNS-mediated effects of insulinare dominant and lower EGP. With an arterial pancreaticclamp, however, the portal-to-peripheral glucagon gradi-ent (23) is also lost, which may have minimized theeffects of relative hepatic insulin deficiency on EGP. Con-sistent with the hypothesis that only under conditionsof relative hepatic insulin deficiency can CNS insulinlower EGP, in experiments carried out using dogs, whena pancreatic clamp was instituted with a normal portal-peripheral insulin concentration gradient, CNS insulindelivery augmented glycogen synthesis but had no effecton EGP 4 h after administration (8). In this study, however,the CNS–to–non-CNS insulin gradient was not maintained.Blockade of hypothalamic insulin action in the settingof physiological hyperinsulinemia and portal-peripheral

772 INI and Endogenous Glucose Production Diabetes Volume 64, March 2015

insulin concentration gradient blunted induction of glu-cokinase gene transcription and abrogated the inhibitionof glycogen synthase 3b transcription but caused no netchange in EGP (24). This effect was seen with a similarphysiological increase in CNS, liver, and peripheral circu-lation. Although it is possible some differences might beascribed to interspecies effects, the presence of a portal-peripheral insulin gradient likely abrogated the effect ofCNS insulin action. Intriguingly, there was a significantreduction in mRNA levels without significant changes inprotein expression of gluconeogenic enzymes (8). Whethera study of longer duration, such as our 6-h study, wouldhave affected EGP remains unclear.

Whether INI would lower plasma glucose in a morephysiological setting, in the absence of an arterial pancreaticclamp and presence of a normal portal-peripheral insulinconcentration gradient, over the time course of the currentstudy (between 180 and 360 min after administration)is currently not known. Previous studies using INI in theabsence of a pancreatic clamp reported a very modest(;5%) reduction in plasma glucose concentration within2 h of administration, which may reflect insulin action dueto peripheral spillover (11,14). Notwithstanding interspe-cies differences and differing assays, based on hyperinsuli-nemic clamp studies of dogs with measurement of plasmaand CSF insulin (25), as well as CSF insulin concentrationafter administration of 40 IU of INI in humans (12), it islikely that 40 IU of insulin causes a supraphysiological in-crease in CSF insulin concentration. In a previous studyassessing the effects of activation of extra-pancreaticKATP channels with an arterial pancreatic clamp, a physio-logical increase in insulin concentrations in the placebogroup did not alter EGP, although pharmacological activa-tion of KATP channels did reduce EGP (16). Despite thenonphysiological aspects of the study discussed above, wedemonstrated for the first time that nasally administeredinsulin decreases EGP in humans compared with INP in thepresence of similar venous insulin concentrations.

Resistance to direct insulin action in the liver andincreased HGP are hallmarks of type 2 diabetes (1). Cur-rent treatment modalities for type 2 diabetes have poten-tial side effects, including weight gain with subcutaneousinsulin and sulphonylureas, which can potentially exacer-bate hepatic insulin resistance (26). Whether INI canacutely lower EGP in individuals with insulin resistanceand type 2 diabetes and whether chronic treatment withINI affects glycemic control remain to be determined. Anadditional potential advantage of INI is that, unlike sub-cutaneous insulin, it is less likely to cause weight gain.Animal studies have suggested that CNS insulin actionreduces appetite and results in weight loss (27). In humanstudies of INI, an acute reduction in appetite was reportedafter a single dose in men (10), with a reduction in post-prandial satiety and intake of palatable food in women(11). In addition, 8 weeks of INI treatment in slim healthyvolunteers modestly reduced fat mass and body weight inmen; modest weight gain occurred in women, which is

thought to be due to an increase in extracellular water,but, importantly, there was no change in body fat (19).It must be noted, however, that in rodent models of in-sulin resistance induced by a high-fat diet, CNS-mediatedeffects of insulin are blunted (28), suggesting the presenceof hypothalamic insulin resistance. Whether a similar phe-nomenon occurs in obese insulin resistant humans remainsto be determined.

In conclusion, we showed that INI, at a dose that isknown to increase CSF insulin concentration, lowers EGPunder conditions of experimental portal hypoinsulinemiacompared with INP. This effect occurs despite similarvenous insulin concentrations between treatments. Addi-tional studies are needed to evaluate whether this path-way is amenable to therapeutic manipulation in insulinresistance and type 2 diabetes.

Acknowledgments. The authors are indebted to Brenda Hughes andPatricia Harley (University Health Network) for their nursing assistance in con-ducting the clinical protocol and Linda Szeto (University Health Network) fortechnical assistance.Funding. S.D. and C.M. are recipients of postdoctoral fellowship awards fromthe Banting and Best Diabetes Centre, University of Toronto, and S.D. is therecipient of a Focus on Stroke 12 Fellowship Award from the Heart and StrokeFoundation of Canada. K.K. is funded by a Juvenile Diabetes Research Foundation-Canadian Clinical Trials Network postdoctoral fellowship. G.F.L. holds the SunLife Financial Chair in Diabetes and the Drucker Family Chair in DiabetesResearch.Duality of Interest. This study was funded by Eli Lilly Canada (grant no. F3Z-CA–0082 to G.F.L.). G.F.L. has served on advisory boards to Eli Lilly Canada. No otherpotential conflicts of interest relevant to this article have been reported.Author Contributions. S.D., C.X., C.M., and G.F.L. designed the studyand interpreted data. S.D., C.X., C.M., and K.K. acquired and analyzed data. S.D.and G.F.L. wrote the manuscript. G.F.L. obtained funding and supervised thestudy. All authors edited the manuscript. G.F.L. is the guarantor of this work and,as such, had full access to all the data in the study and takes responsibility forthe integrity of the data and the accuracy of the data analysis.

References1. Lin HV, Accili D. Hormonal regulation of hepatic glucose production inhealth and disease. Cell Metab 2011;14:9–192. Kawamori D, Kurpad AJ, Hu J, et al. Insulin signaling in alpha cellsmodulates glucagon secretion in vivo. Cell Metab 2009;9:350–3613. Lewis GF, Vranic M, Giacca A. Role of free fatty acids and glucagon in theperipheral effect of insulin on glucose production in humans. Am J Physiol 1998;275:E177–E1864. Lewis GF, Vranic M, Harley P, Giacca A. Fatty acids mediate the acuteextrahepatic effects of insulin on hepatic glucose production in humans. Diabetes1997;46:1111–11195. Obici S, Zhang BB, Karkanias G, Rossetti L. Hypothalamic insulin signalingis required for inhibition of glucose production. Nat Med 2002;8:1376–13826. Pocai A, Lam TK, Gutierrez-Juarez R, et al. Hypothalamic K(ATP) channelscontrol hepatic glucose production. Nature 2005;434:1026–10317. Filippi BM, Yang CS, Tang C, Lam TK. Insulin activates Erk1/2 signalingin the dorsal vagal complex to inhibit glucose production. Cell Metab 2012;16:500–5108. Ramnanan CJ, Saraswathi V, Smith MS, et al. Brain insulin actionaugments hepatic glycogen synthesis without suppressing glucose production orgluconeogenesis in dogs. J Clin Invest 2011;121:3713–3723

diabetes.diabetesjournals.org Dash and Associates 773

9. Guthoff M, Grichisch Y, Canova C, et al. Insulin modulates food-relatedactivity in the central nervous system. J Clin Endocrinol Metab 2010;95:748–75510. Benedict C, Kern W, Schultes B, Born J, Hallschmid M. Differentialsensitivity of men and women to anorexigenic and memory-improving effects ofintranasal insulin. J Clin Endocrinol Metab 2008;93:1339–134411. Hallschmid M, Higgs S, Thienel M, Ott V, Lehnert H. Postprandial ad-ministration of intranasal insulin intensifies satiety and reduces intake of palat-able snacks in women. Diabetes 2012;61:782–78912. Born J, Lange T, Kern W, McGregor GP, Bickel U, Fehm HL. Sniffingneuropeptides: a transnasal approach to the human brain. Nat Neurosci 2002;5:514–51613. Brooking J, Davis SS, Illum L. Transport of nanoparticles across the ratnasal mucosa. J Drug Target 2001;9:267–27914. Heni M, Kullmann S, Ketterer C, et al. Nasal insulin changes peripheralinsulin sensitivity simultaneously with altered activity in homeostatic and reward-related human brain regions. Diabetologia 2012;55:1773–178215. Heni M, Wagner R, Kullmann S, et al. Central insulin administrationimproves whole-body insulin sensitivity via hypothalamus and parasympatheticoutputs in men. Diabetes 2014;63:4083–408816. Kishore P, Boucai L, Zhang K, et al. Activation of K(ATP) channelssuppresses glucose production in humans. J Clin Invest 2011;121:4916–492017. Vella A, Rizza RA. Application of isotopic techniques using constantspecific activity or enrichment to the study of carbohydrate metabolism. Diabetes2009;58:2168–217418. Iwen KA, Scherer T, Heni M, et al. Intranasal insulin suppresses systemicbut not subcutaneous lipolysis in healthy humans. J Clin Endocrinol Metab 2014;99:E246–E251

19. Hallschmid M, Benedict C, Schultes B, Fehm HL, Born J, Kern W.Intranasal insulin reduces body fat in men but not in women. Diabetes 2004;53:3024–302920. Edgerton DS, Cherrington AD, Neal DW, et al. Inhaled insulin is asso-ciated with prolonged enhancement of glucose disposal in muscle and liver in thecanine. J Pharmacol Exp Ther 2009;328:970–97521. Everard ML, Hardy JG, Milner AD. Comparison of nebulised aerosoldeposition in the lungs of healthy adults following oral and nasal inhalation.Thorax 1993;48:1045–104622. Edgerton DS, Lautz M, Scott M, et al. Insulin’s direct effects on the liverdominate the control of hepatic glucose production. J Clin Invest 2006;116:521–52723. Jaspan JB, Ruddick J, Rayfield E. Transhepatic glucagon gradients inman: evidence for glucagon extraction by human liver. J Clin Endocrinol Metab1984;58:287–29224. Ramnanan CJ, Kraft G, Smith MS, et al. Interaction between the centraland peripheral effects of insulin in controlling hepatic glucose metabolism in theconscious dog. Diabetes 2013;62:74–8425. Schwartz MW, Sipols A, Kahn SE, et al. Kinetics and specificity of insulinuptake from plasma into cerebrospinal fluid. Am J Physiol 1990;259:E378–E38326. Scheen AJ, Van Gaal LF. Combating the dual burden: therapeutic targetingof common pathways in obesity and type 2 diabetes. Lancet Diabetes Endocrinol2014;2:911–92227. Woods SC, Lotter EC, McKay LD, Porte D Jr. Chronic intracerebroventricularinfusion of insulin reduces food intake and body weight of baboons. Nature 1979;282:503–50528. Benoit SC, Kemp CJ, Elias CF, et al. Palmitic acid mediates hypothalamicinsulin resistance by altering PKC-theta subcellular localization in rodents. J ClinInvest 2009;119:2577–2589

774 INI and Endogenous Glucose Production Diabetes Volume 64, March 2015