Page 1

Intranasal insulin prevents cognitive decline,

cerebral atrophy and white matter changes in

murine type I diabetic encephalopathy

George J. Francis,1 Jose A. Martinez,1 Wei Q. Liu,1 Kevin Xu,1 Amit Ayer,1 Jared Fine,2 Ursula I. Tuor,1

Gordon Glazner,3 Leah R. Hanson,2 William H. Frey II2 and Cory Toth1

1Department of Clinical Neurosciences and the Hotchkiss Brain Institute, University of Calgary, Calgary, Alberta,

Canada, 2Alzheimer’s Research Center at Regions Hospital, HealthPartners Research Foundation, St. Paul, MN, USA

and 3Department of Pharmacology and Therapeutics, Division of Neurodegenerative Disorders, University of Manitoba,

St. Boniface Hospital Research Centre, Winnipeg, Manitoba, Canada

Correspondence to: Dr C. Toth, University of Calgary, Department of Clinical Neurosciences, Room 155, 3330 Hospital Drive,

N.W., Calgary, Alberta, Canada T2N 4N1

E-mail: corytoth@shaw.ca

Insulin deficiency in type I diabetes may lead to cognitive impairment, cerebral atrophy and white matter

abnormalities. We studied the impact of a novel delivery system using intranasal insulin (I-I) in a mouse

model of type I diabetes (streptozotocin-induced) for direct targeting of pathological and cognitive deficits

while avoiding potential adverse systemic effects. Daily I-I, subcutaneous insulin (S-I) or placebo in separate

cohorts of diabetic and non-diabetic CD1 mice were delivered over 8 months of life. Radio-labelled insulin

delivery revealed that I-I delivered more rapid and substantial insulin levels within the cerebrum with less

systemic insulin detection when compared with S-I. I-I delivery slowed development of cognitive decline

within weekly cognitive/behavioural testing, ameliorated monthly magnetic resonance imaging abnormalities,

prevented quantitative morphological abnormalities in cerebrum, improved mouse mortality and reversed

diabetes-mediated declines in mRNA and protein for phosphoinositide 3-kinase (PI3K)/Akt and for protein

levels of the transcription factors cyclic AMP response element binding protein (CREB) and glycogen synthase

kinase 3b (GSK-3b) within different cerebral regions. Although the murine diabetic brain was not subject to

cellular loss, a diabetes-mediated loss of protein and mRNA for the synaptic elements synaptophysin and

choline acetyltransferase was prevented with I-I delivery. As a mechanism of delivery, I-I accesses the brain

readily and slows the development of diabetes-induced brain changes as compared to S-I delivery. This therapy

and delivery mode, available in humans, may be of clinical utility for the prevention of pathological changes in

the diabetic human brain.

Keywords: diabetes; insulin; leukoencephalopathy; white matter abnormalities; brain atrophy; cognitive decline

Abbreviations: APP = amyloid precursor protein; CSF = cerebrospinal fluid; DW = diffusion-weighted;

EMSA = electrophoretic mobility shift assay; IGF-1 = insulin-like growth factor; IR = insulin receptor;

NGF = nerve growth factor; SNMTS = spatial non-matching-to-sample; WMAs = white matter abnormalities

Received May 27, 2008. Revised October 3, 2008. Accepted October 8, 2008. Advance Access publication November 16, 2008

Introduction

Diabetes has been associated with cognitive dysfunction, an

elevated risk of dementia, cerebral atrophy and presence of

heightened white matter abnormalities (WMAs). Diabetes-

associated cognitive dysfunction, first described nearly a

century ago (Miles and Root, 1922), occurs in both type of

diabetes. In type I diabetes, impaired learning, memory,

problem solving skills, and intellectual development have

been described (Ryan, 1988; Ryan et al., 1993; Ryan and

Williams, 1993; McCarthy et al., 2002; Schoenle et al.,

2002). For patients with type I diabetes, the greatest impact

of diabetes upon brain structure and function seems to

occur at the extremes of age, with little observable effect

during the middle adult years (Wessels et al., 2007, 2008;

Biessels et al., 2008; Kloppenborg et al., 2008). Also,

cognitive dysfunction does not appear to relate to

doi:10.1093/brain/awn288

Brain (2008),131, 3311^3334

ß The Author (2008). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: journals.permissions@oxfordjournals.org

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

hypoglycaemic episodes in diabetes (Kramer et al., 1998;

Schoenle et al., 2002; Jacobson et al., 2007). Similar

cognitive deficits occur in type II diabetic adult patients,

with impaired performance on abstract reasoning and

complex psychomotor functioning (Reaven et al., 1990;

Strachan et al., 1997; Ryan and Geckle, 2000). Diabetes-

mediated cognitive changes seem to occur at the two

extremes of life: either in childhood (Northam et al., 2001;

Schoenle et al., 2002; Fox et al., 2003) or during later stages

of life (Stewart and Liolitsa, 1999; Awad et al., 2004; van

Harten et al., 2006) when neurodegenerative processes

may initiate (Biessels et al., 2008). There are less clear

indications that diabetes impacts upon cognitive function-

ing in middle-age adults (Stewart and Liolitsa, 1999; Awad

et al., 2004; Jacobson et al., 2007; Weinger et al., 2008).

In animal models of diabetes, cognitive dysfunction has

been demonstrated by impaired performances in the Morris

water maze by streptozotocin (STZ)-induced diabetic rats

(Biessels et al., 1996, 1998). Both pre- and post-synaptic

deficits have been associated with impaired long-term

potentiation in the diabetic hippocampus (Biessels et al.,

1996; Kamal et al., 2006). Short-term replacement of insulin

in STZ-treated rats from the onset of diabetes prevents

cognitive decline and protects against hippocampal poten-

tiation deficits, but cannot reverse these electrophysiological

changes (Biessels et al., 1998). Long-term protection against

development of diabetic encephalopathy via locally deliv-

ered insulin has not yet been attempted.

Structural defects within the diabetes-exposed brain

include cerebral atrophy identified with neuroimaging

techniques (Schmidt et al., 2004; Manschot et al., 2006;

Musen et al., 2006; Ikram et al., 2008; Last et al., 2007).

Cerebral atrophy, possibly acting in concert with WMA, is

associated with cognitive decline (Whitman et al., 2001;

Manschot et al., 2006). Diabetes is a risk factor for WMA

presence in some studies (Pantoni and Garcia, 1997;

Murray et al., 2005; Akisaki et al., 2006; Manschot et al.,

2006; Musen et al., 2006; van Harten et al., 2007), but not

in others (Schmidt et al., 2004). By themselves, WMAs in

humans are a risk factor for stroke (Knopman et al., 2001),

cognitive deficits (Pantoni et al., 2007) and abnormalities in

gait associated with falling (Schwartz et al., 2008). The

presence of cerebral atrophy or WMA has been linked to

cognitive dysfunction in a number of human studies

(Manschot et al., 2006; Verdelho et al., 2007). Rodent

models of diabetes have also demonstrated cerebral atrophy

(Lupien et al., 2006; Toth et al., 2006) and WMAs (Toth

et al., 2006) due to long-term diabetes. Such pathological

changes have been associated with cognitive decline over

time in mouse models of diabetes (Toth et al., 2006) and in

diabetic rats where hippocampal electrophysiological

changes were present (Biessels et al., 1996; Kamal et al.,

2000). One pathogenic factor related to presence of brain

atrophy and WMA in experimental diabetes is the presence

of the receptor for advanced glycation end products

(RAGE) (Toth et al., 2006). However, another important

factor may be relative impaired insulin levels and activity in

the brain exposed to diabetes (Li et al., 2005; Haan, 2006),

as described in the human Alzheimer’s disease brain (Craft

et al., 1998; Hoyer, 2004). Insulin receptors (IRs) are

present at central neurons, synapses, and upon glia (Adamo

et al., 1989; Unger et al., 1989; Wickelgren, 1998; Abbott

et al., 1999; Zhao et al., 2004a); therefore, it is assumed that

insulin and its signalling play an important role in neur-

onal, glial, and overall cognitive and memory functioning.

Insulin replacement within the central nervous system may

prevent or even reverse such diabetes-associated changes,

although its systemic delivery is complicated by hypogly-

caemia. Also, potentially impaired function of the blood–

brain barrier may prevent insulin transport from achieving

sufficient cerebral levels (Cohen, 1993; Mooradian, 1997;

Hattori et al., 2000; Kaiyala et al., 2000). An alternative

method of delivering insulin to the brain could be

instrumental

in

preventing

diabetes-accelerated

neurodegeneration.

Intranasal administration permits insulin or similar

peptides to bypass the periphery and the blood–brain

barrier (Dhanda et al., 2005), reaching the brain and

entering the cerebrospinal fluid (CSF) within minutes.

Proteins with size of up to 20kDa, including insulin,

insulin-like growth factor (IGF)-1 and nerve growth factor

(NGF), have been successfully delivered to the brain using

this method (Chen et al., 1998; Liu et al., 2004; Ross et al.,

2004; Thorne et al., 2004; Reger et al., 2006, 2008; Vig et al.,

2006). Transport of molecules delivered intranasally occurs

through extracellular bulk flow transport along olfactory

and trigeminal perivascular channels, as well as possibly

through axonal transport pathways (Benedict et al., 2004;

Thorne et al., 2004; Reger et al., 2006). This technique

permits targeted delivery to the brain to assess the direct

effect of insulin upon its receptors without significant

changes in plasma insulin or glycaemic levels (Patti et al.,

1995; White and Yenush, 1998; Leinninger et al., 2004;

Reger et al., 2006, 2008; Yu et al., 2006).

We hypothesized that long-term replacement of insulin

within the experimental type I diabetic mouse brain could

slow or prevent such changes. Insulin’s ligation to highly

expressed IR in brain regions including the hippocampus

and upon synapses (Abbott et al., 1999), may promote

learning and memory (Zhao et al., 2004a). We used daily

intranasal delivery of insulin over a life span while per-

forming behavioural and neuroimaging measures to

monitor progression. These studies were also performed

to assist in delineation of insulin’s trophic and anti-

hyperglycaemic effects upon development of diabetic

encephalopathy. Given the absence of neuronal loss in

prior studies of the diabetic murine brain (Toth et al.,

2006), we also examined for evidence of synaptic loss, as

well as the previously detected myelin loss. We postulated

that intranasal insulin delivery would sustain the diabetic

nervous system, potentially limiting cognitive decline, brain

3312

Brain (2008), 131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

atrophy and WMAs, while limiting serious systemic side-

effects that could occur with systemic insulin delivery.

Materials and Methods

Animals

We studied a total of 484 male Swiss Webster wild-type (wt) mice

with initial weight of 20–30g, in strict pathogen-free plastic

sawdust-covered cages with a normal light–dark cycle and free

access to mouse chow and water. All protocols were reviewed and

approved by the University of Calgary Animal Care Committee

using the Canadian Council of Animal Care guidelines. Mice were

anaesthetized with pentobarbital (60mg/kg) prior to all proce-

dures. At the age of 1 month, 304 mice were injected with

streptozotocin (Sigma, St. Louis, MO) intraperitoneally with once

daily doses with each of 60mg/kg, 50mg/kg and then 40mg/kg

over three consecutive days, while the remaining 180 mice were

injected with volume-equated placebo carrier (sodium citrate) for

3 consecutive days. One hundred and forty-four mice injected

with STZ and 80 mice injected with carrier were held aside for

morphological studies to be performed at 1, 3 and 5 months after

injections; the remaining mice were followed for the entire length

of the study (8 months of diabetes or equivalent for carrier-

injected mice) as mortality permitted. Monthly weights were

obtained and monthly whole-blood glucose measurements were

performed using the tail vein and a blood glucometer, (OneTouch

Ultra Meter, LifeScan Canada, Burnaby, BC, Canada) with hyper-

glycaemia verified 1 week after STZ injection; a fasting whole-

blood glucose level of 516mmol/l (normal 5–8mmol/l) was our

criterion for experimental diabetes. In all cases, those mice that

did not develop diabetes as defined above following STZ injections

were excluded from further assessment. Animals were inspected

twice daily, and examined for signs of depressed level of con-

sciousness, ataxia or general malaise. When such signs were

identified, whole-blood glucose testing was performed, with a

measurement of53.5mmol/l defined to represent hypoglycaemia.

No intervention was performed at any time with regards to

additional insulin, glucose or fluid delivery. In situations where

the mouse was obviously ill, euthanasia was performed. In circu-

mstances where severe hyperglycaemia was found (433 mmol/l) in

an ill mouse, euthanasia was again performed.

We studied cohorts with a maximum of eight mice in each

group initially due to resource limitations. After the initial cohorts

containing eight mice each were studied, a second cohort was used

to obtain additional mouse data for mouse cohorts with higher

levels of mortality. For any animal that experienced mortality after

the 20-week point of the cognitive studies, their data were carried

through using the last obtainable data point.

Intranasal insulin or saline delivery

125I-labelled I-I and subcutaneous insulin (S-I) administration

were performed at the University of Minnesota for determination

of targeting of insulin delivery methods. This procedure was

approved by the institutional animal care and use committee at

Regions Hospital. Prior to experimentation, 21 non-diabetic

animals were acclimated to handling for awake intranasal delivery

over $2 weeks. 125I-labelled I-I was provided to 12 Swiss Webster

mice (male, 6–8 weeks, Charles River) and 125I-labelled sub-

cutaneous insulin S-I was provided to nine mice under

pentobarbital anaesthesia (60mg/kg). Insulin (Humulin R, Eli

Lilly, Toronto, Canada) with an initial concentration of 100U/ml

or 4033.98mg/ml was dissolved in PBS and custom-labelled with

125I (GE Healthcare, Piscataway, NJ, USA). Synthesized radio-

labelled insulin solution contained 344.3 mCi/ug. 125I-labelled I-I

delivery was performed in a fume hood behind a lead-

impregnated shield, with anaesthetized mice placed supine. A

mixture of 125I insulin (15.8 mCr) and unlabeled insulin (3.3 mg)

were administered as I-I or S-I. 125I I-I was delivered over

alternating nares as eight 3-ml drops with an Eppendorf pipetter

every 2min, for a total volume of 24 ml. This schedule of delivery

has been modified from a previously used method in rats to

quantify radiolabelled delivery of molecules within the Frey

laboratory (Thorne et al., 2004). For subcutaneous delivery, 125I

S-I was delivered with a single subcutaneous injection of 24 ml in

a fume hood behind a lead-impregnated shield. Each desired dose

contained a calculated radioactive dose of 30 mCi.

At each of 1, 2 and 6h after initiating 125I I-I or S-I delivery,

cardiocentesis was performed for blood extraction, followed by

performance of euthanasia via transcardial perfusion using 120ml

of 4% paraformaldehyde while the mouse was maintained under

anaesthesia. To quantify 125I distribution, blood, urine, lymphatic

and visceral organ structures, as well as portions of the central and

peripheral nervous systems were harvested. Gamma signal was

recorded for each body region with autoradiographic imaging

using a phosphor screen. Concentrations of 125I insulin were

calculated based upon the gamma counting data, tissue weight,

specific activity of the insulin administered and standards

measured. Results were studied for penetration into peripheral

nervous system tissues (reported elsewhere) and central nervous

system tissues.

Daily I-I (Humulin R, Eli Lilly, Toronto, ON) or intranasal

saline (I-S) was administered to both diabetic and non-diabetic

male Swiss Webster mice over 8 months of diabetes. A total of

24 ml containing either a total of 0.87U of insulin or 0.9% saline

only was provided as four 6-ml drops by an Eppendorf pipetter

over alternating nares every minute while each mouse was held in

supine position with neck in extension. Daily S-I (0.87U/d,

Humulin R, Eli Lilly, Toronto, ON) and subcutaneous saline (S-S)

were also administered to either diabetic or non-diabetic male

Swiss Webster mice. All therapies began immediately after

confirmation of presence of diabetes for each cohort. In the first

week, daily glucometer testing was performed for all mice,

followed by once-a-month testing. In this work, mice with

diabetes were indicated with a ‘D’, while mice without diabetes

(control mice) are indicated with a ‘C’. Delivery of subcutaneous

saline is indicated as ‘S-S’, subcutaneous insulin as ‘S-I’, intranasal

saline as ‘I-S’ and intranasal insulin as ‘I-I’.

We attempted to use other control groups, but their utility was

limited in each case. First, we attempted to use subcutaneous

insulin via a sliding scale approach in six diabetic mice in order to

maintain normal or mildly high glycaemia levels. This approach

required daily checks of whole-blood glucose using a tail vein;

over the span of 1 month, all of the six diabetic mice developed

infection over the tail, leading to amputation in three mice (50%)

and was probable cause of death in two mice (33%). During the

course of an 8-month-long study, the morbidity associated with

this procedure would be unacceptable and confounding. We also

attempted to maintain a protected venous catheter in the tail vein

for obtaining whole-blood glucose, but this was associated with

auto-removal of the catheter and failure at re-insertions due to

fibrosis of tail tissues. We also studied six diabetic mice receiving

Intranasal insulin and diabetic encephalopathy

Brain (2008), 131, 3311^3334

3313

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

a half dose of the desired I-I dose (0.43U/d) over 1 month. One

mouse (17%) died of confirmed hypoglycaemia, with the other

five mice surprisingly had no statistical difference in glycaemic

levels when compared to a complementary group of six diabetic

mice receiving subcutaneous placebo injections. Therefore, we

selected the S-I dose to be equivalent to the I-I dose for the

cohorts studied.

Behavioural testing

A minimum of eight mice in each cohort [32 diabetic (D I-I, D I-

S, D S-I and D S-S) and 32 control mice (C I-I, C I-S, C S-I and

C S-S)] had cognitive behavioural testing performed once weekly

for evaluation of procedural, visuospatial and recognition

memory. All behavioural tests were conducted under regular

light between 09:00 and 15:00h after fasting overnight from 19:00

onwards. Every mouse was trained in each test paradigm for

3 weeks before initiation of test recording and prior to diabetes

initiation, beginning after 2 weeks of diabetes at 1.5 months of

age. Behavioural equipment was maintained in the identical

position and climate on each occasion and was recorded by the

same observers in order to ensure stability of distant navigational

cues provided by objects around the testing. Testing always

occurred in the order of Holeboard test, Radial arm test, Object

Recognition test (each performed while fasting) followed by

feeding and then the Morris Water-Maze test 1h later. While the

Holeboard and Radial Arm tests evaluate spatial information

processing and memory, the Morris Water Maze also evaluates

procedural memory and aversive motivation. In contrast, the

Object Recognition test evaluates novelty seeking and exploratory

behaviour. Each mouse performed one trial of each test per week.

Due to the possibility of motor limitations confounding the results

of cognitive testing, concurrent testing of linear swim speed and

linear running speed were performed monthly; once measurable

differences in motor function occurred between interventional

groups in swim or run speed, cognitive testing was discontinued.

The Radial Arm test maze consists of a central platform (35cm

in diameter) and eight arms (each 76cm long and 12cm wide)

constructed of black plastic projecting radially from the platform

with adjacent arms separated by 45 (Schwabe et al., 2006). A food

reward (Cheerio) is placed in the same arm each occasion at 180

from the starting point of the mouse placed in the middle of the

central platform. A mouse was recorded as entering an arm when

it passed the midpoint of the arm in a centrifugal fashion. Each

trial ended when the reward was collected or when 720s expired.

Variables recorded during each test were latency for collecting

reward and the number of errors made, defined as a re-entry into

an arm previously entered, classified as a reference memory error.

The Holeboard test was modified from a previous reported

design (File and Wardill, 1975), and was composed of a

rectangular open field (60 Â 90cm) made of opaque white acrylic

surrounded by opaque walls 60cm high. Eight holes (2.5cm

diameter) were placed in two lines of four, equidistant from each

other and from the walls. Each mouse was started in the same

corner while the food reward (Cheerio) was placed in the same

hole (second hole in the far row, kitty-corner from the starting

point). The latency to collect the reward and the number of times

the mouse placed its head in each individual hole was recorded,

with errors defined as repeat visit of a previously visited hole,

classified as a reference memory error.

The Object Recognition task was performed in an open wooden

box (60 Â 60 Â 60cm) with unique objects to be discriminated

constructed from children’s blocks. On the day of the test, at the

first 2-min sample trial (T1), two identical objects (termed as

sample objects) were presented in two corners of the box. Then, in

the second 2-min choice trial (T2) performed 30min later, one of

the objects presented in T1 was replaced by a new object. Objects

were cleaned between trials and between mouse to prevent the

possibility of scent traces forming an olfactory cue. The time (in

seconds) as well as the number of visits taken by mice in exploring

objects in the two trials were recorded, with exploration

considered as directing the nose to the object at a distance

42m and/or touching it with the nose. This paradigm has been

termed delayed ‘spatial non-matching-to-sample’ (SNMTS) testing

(Rothblat and Kromer, 1991).

A mouse-adapted Morris water-maze task (Morris, 1984;

Crawley, 2000; Whishaw and Kolb, 2005) was performed after

feeding and used a solid-coloured circular pool 88cm in diameter

and 20cm in height filled with water at 25 C. The position of the

10cm radius hidden platform remained fixed for all testing over

the entire study period, and each mouse was placed at an identical

starting position opposite the hidden platform. Animals were left

to swim until either they located the platform, climbing upon it

and staying for at least 2s, or when 300s elapsed. Post-testing,

mice were placed under a heating lamp to warm. Variables

recorded during each test were latency to reach the platform

(escape latency) and the fraction of time spent within the

hemisphere of the pool containing the platform (thigmotaxis).

Magnetic resonance imaging

At each month of diabetes, four mice from each cohort [16

diabetic (D I-I, D I-S, D S-I and D S-S) and 16 control mice

(C I-I, C I-S, C S-I and C S-S)] underwent magnetic resonance

(MR) scanning at the Experimental Imaging Centre at the

University of Calgary. MR images were obtained in animals

anaesthetized by mask with isoflurane using a quadrature volume

coil and a Bruker 9.4Tesla MR imaging system. Respiration and

temperature were monitored and inner core temperature was

maintained to be within 36–37 C with a heated air feedback

system. The head was restrained using ear pins. Three different

sets of MR scans were performed using sequences that acquired

T1-weighted images, diffusion-weighted (DW) images for calculat-

ing an apparent diffusion coefficient of water (ADC) map and T2-

weighted images for determining T2 maps within a total of 24

slices through the cerebrum. T1-weighted images were acquired

using a spin echo sequence with a repetition time (TR) of 500ms

and an echo time (TE) of 8ms. Diffusion-weighted (DW) images

were acquired using a spin echo sequence with TR/TE=1200/

49ms and b values of 46 and 767s/mm2, respectively. T2 maps

were obtained from multi-spin-echo images at TR=1200ms, 12

echoes and TE=12.5ms. The field of view was 2 Â 2cm, with an

acquisition matrix of 256 Â256 and a slice thickness of 0.75mm.

MR images of brain were analysed using locally available software

(Marevisi, IBD) by an observer blinded to the treatment group.

Calculation of T2 values, perfusion values, and ADC values within

brain regions of interest were performed bilaterally for each of the

control and diabetic animals. Apparent visualized abnormalities in

T1- and T2-weighted images were also recorded for each animal.

In addition, MR images were used to calculate brain widths at

pre-determined anatomical landmarks as well as to measure

3314

Brain (2008), 131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

overall brain volume. Volumetric brain measurements were

calculated as a summation of cross-sectional areas for each slice

multiplied by the thickness of the MR slices.

Tissue harvesting

Prior to sacrifice after 1, 3, 5, 7 or 9 months of diabetes, animal

weights and blood glucoses were determined. A total of 0.5ml of

whole intracardiac blood was obtained for glycated haemoglobin

measurements to be performed with affinity chromatography

(Procedure 422; Sigma Diagnostics). Euthanasia for animals was

performed with an overdose of pentobarbital intraperitoneally,

followed by harvesting of brain tissues, placement of half of all

brain tissues in 2% formaldehyde and embedding in paraffin. For

these specimens, 10 mm sections were cut for morphological and

immunohistochemical studies. The other half of mouse brain

tissues were placed in Trizol reagent (Life Technologies Inc.,

Rockville MD, USA) or liquid nitrogen for quantitative real-time

reverse transcriptase polymerase chain reaction (qRT-PCR) and

Western blot studies respectively with storage at –80 C for a

maximum of 1 month.

Brain sectioning and staining

Degrees of myelination were determined by staining paraffin brain

sections for either Luxol Fast Blue (LFB) or myelin basic protein

(MBP) (1:100, Stemcell Technologies Inc., Vancouver). For LFB

staining, slides were de-waxed and stained in LFB overnight at

37 C followed by alcohol washing and subsequent differentiation

of the stain with 0.05% lithium carbonate and then alcohol. Slides

used for MBP detection were incubated in methanol for 20min

and washed in phosphate-buffered solution (PBS), then incubated

in Triton-X for 30min before blocking with 10% normal bovine

serum for 1h. Following PBS washes, slides were incubated with

mouse anti-MBP overnight followed by incubation with the

secondary antibody (bovine anti-mouse IgG Cy3, 1:100 Zymed

Inc., San Francisco) for 1h.

Brain sections stained for LFB were chosen to reflect pre-

defined regions of interest (Appendix 1), representing those

regions known to be abnormal within diabetic human brains, as

well as cortical and subcortical regions important in memory and

cognition (Munoz et al., 1993; de Groot et al., 2000; Tullberg

et al., 2004; Toth et al., 2006). Assessment of MBP labelling was

performed using Image Pro Plus software (Image Pro Plus 5.0,

MediaCybernetics, Silver Spring, MD) in order to measure the

optical density or degree of immunofluorescence within each brain

region of interest. LFB staining was assessed using optical density

measurements with Photoshop software for quantification of blue.

All assessments of brain sections were performed by an

investigator blinded to the experimental conditions.

Additional immunohistochemistry was performed to identify

neurons with microtubule-associated protein (MAP)-2, and to

identify oligodendrocytes with PDGFRa immunohistochemistry.

Background staining was blocked using 1% bovine serum albumin

for 1h, and then slides were stained with monoclonal mouse

anti-MAP-2 (1:100, Abcam, Cambridge, MA) or PDGFRa

(1:100, Abcam, Cambridge, MA) for neuron and oligodendro-

cyte detection, respectively. The secondary antibody used was

anti-mouse IgG Cy3 (1:100, Zymed Inc., San Francisco) in both

cases.

Detailed counts of both neuronal and oligodendroglial cell

numbers within grey matter areas of interest (Appendix 1) were

performed using standard unbiased stereological methods for

those slides stained with the neural marker MAP-2 (Gundersen

et al., 1988; Toth et al., 2006) and the oligodendroglial marker

PDGFRa. Volume and total cell number were performed with the

examiner blinded to the condition and treatment of each mouse

brain. Twelve sampled areas were counted for each brain region in

each brain, with neurons distinguished based upon nuclear size

and appearance.

Quantitative real-time PCR

Total RNA was extracted from brain regions stored in Trizol

reagent (Life Technologies Inc., Rockville, MD). Total RNA (1 mg)

was processed directly to cDNA synthesis using the Taq-

Man\Reverse Transcription Reagents kit (Applied Biosystems).

All PCR primers and TaqMan probes were designed using soft-

ware PrimerExpress (Applied Biosystem) and published sequence

data from the NCBI database. PI3K primer sequences were: for-

ward, 50-AACCCGGCACTGTGCATAAA-30; reverse 50-GCCCATT

GGATTAGCATTGATG-30. Akt primer sequences were: forward,

50-TCTGCCCTGGACTACTTGCACT-30, reverse, 50-GCCCGAAGT

CCGTTATCTTGA-30. NFkBp65 primer sequences were: forward,

50-TGTGCGACAAGGTGCAGAAA-30; and reverse, 50-ACAATG

GCCACTTGCCGAT-30. b-actin primer sequences were: forward,

50-TGTTGTCCCTGTATGCCTCTGGTC-30; reverse, 50-ATGTCA

CGCACGATTTCCCTCTCTC-30. SYP primer sequences were: for-

ward, 50-AAAGGCCTGTCCGATGTGAAG-30; reverse, 50-TCCC

TCAGTTCCTTGCATGTG -30. ChAT primer sequences were:

forward, 50-CTATGAGAGTGCATCCATCCGC-30; reverse, 50-GG

TCAGTCATGGCTTGCACAA-30.

RT-PCR was performed using SYBR Green dye. All reactions

were performed in triplicate in an ABI PRISM 7000 Sequence

Detection System. Data were calculated by the 2–AACT method

and are presented as the fold induction of mRNA for RAGE in

diabetic tissues normalized to 18S for comparison to non-diabetic

tissues (defined as 1.0-fold).

Western blot

Thalami and sensorimotor cortices from brains for each cohort

were homogenized using a RotorStator Homogenizer in ice-cold

lysis buffer (10% glycerol, 2% SDS, 25mM Tris–HCl, pH 7.4,

Roche Mini-Complete Protease Inhibitors). Samples were then

centrifuged at 10000 g for 15min. Supernatant was stored at

–20 C prior to SDS–PAGE and immunoblotting analysis. Equal

amounts (150 mg) of protein were loaded and samples were

separated by SDS–PAGE using 10% polyacrylamide gels with

800V-h of current applied. Separated proteins were transferred

onto nitrocellulose paper (BioRad) over 16 h at 200 mA in Towbin

transfer buffer (25mM Tris, 192mM glycine, 20% v/v methanol,

0.1% v/v SDS). The blot was blocked for 1h in 7.5% (w/v) milk

(Nestle, Carnation) in TBS [50mM Tris, 137mM NaCl, 51mM

KCL, 0.05% (v/v) Tween-20]. The PI3K and Akt pathway were

investigated with PI3K (1:1000), PKB/Akt (1:1000), pAkt (1:1000).

The nuclear signalling transcription factor NFkB p65 and p50

subunits (1:1000 each) were also examined, and additional

immunohistochemistry was performed for CREB (1:200, Abcam

Inc., Cambridge, MA) and glycogen synthase kinase 3b (GSK3b)

(1:200, Abcam Inc., Cambridge, MA). Quantification of synaptic

presence was performed using anti-synaptophysin (SYP) (1:1000;

Santa Cruz, Santa Cruz, CA, USA, polyclonal) and Anti-

Choline Acetyltransferase (ChAT) (1:500, Abcam, Cambridge,

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3315

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

UK, polyclonal). Identification of proteins for IRb (C-19) (1:500,

Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) and insulin

(1:500, Abcam, Cambridge, UK, polyclonal) were also performed

using Western blotting. For a housekeeping protein, anti-b-actin

(1:100, Biogenesis Ltd. Poole, UK) was applied to separate blots.

Secondary anti-rabbit, anti-mouse or anti-human IgG HRP Linked

antibody (Cell Signaling) was applied at 1:5000 in each case as

appropriate. Signal detection was performed by exposing of the

blot to enhanced chemi-luminescent reagents ECL (Amersham)

for 2min. The blots were subsequently exposed and captured on

Kodak X-OMAT K film. In each case, three blots were performed,

and analysed with Adobe Photoshop (Adobe Photoshop 9.0,

Adobe, San Jose, CA, 2005) for quantification of relative protein

content.

Analysis of Western blots used a ratio of protein of interest to

b-actin protein for each region of brain tissue and each cohort.

Quantification of the luminosity of each identified protein band

was performed using Adobe Photoshop software (Adobe

Photoshop 7.0, Adobe, San Jose, CA, 2002).

Additional immunohistochemistry

Immunohistochemistry was performed using PI3K (1:200,

Santa Cruz Inc., Santa Cruz, CA), PKB/Akt [1:200, anti-protein

kinase B (Akt), Stressgen, Victoria, Canada], pAkt [1:200, anti-

phospho-Akt (Ser473), Cell Signaling Technologies, Danvers, MA]

and the nuclear signalling transcription factor NFkB p65

subunit (1:200, anti-NFkB p65, Santa Cruz, Santa Cruz, CA)

and p50 subunit (1:200, anti-NFkB p50, Santa Cruz, Santa Cruz,

CA). For synaptic identification, anti-synaptophysin (1:200;

Santa Cruz, Santa Cruz, CA, USA, polyclonal) and Anti-

Choline Acetyltransferase (ChAT) (1:100, Abcam, Cambridge,

UK, polyclonal) were used. Identification of IR and insulin

was performed with immunohistochemistry using antibodies

to IRb (C-19) (1:100, Santa Cruz Biotechnology Inc., Santa

Cruz, CA, USA).

Tissue specimens were examined under fluorescence microscopy

(Zeiss Axioskope, Axiovision and Axiocam, Zeiss Canada,

Toronto, Canada) at 400Â and images obtained were examined

based upon brain regions of interest sectioned at 10 mm (Appendix

1). Calculation of the number of immunofluorescent profiles as

well as the relative luminosity was performed using Adobe

Photoshop (Adobe Photoshop 9.0, Adobe, San Jose, CA, 2005).

In grey matter regions, the total numbers of neurons per trans-

verse section, as well as the numbers of neurons with positive

immunolabelling for the above markers, and their potential

nuclear activation, were recorded. Luminosity was classified as

none-low (luminosity value of 0–150), moderate (150–250) or

high (4250) using Adobe Photoshop software (scale of 0–255 with

arbitrary units). An additional measurement of neuronal nuclear

immunolabelling for pAkt and NFkB was also performed using a

pre-determined luminosity measurement threshold of 150 (no

units), below which negative nuclear reactivity was assigned for

both neurons and glia. All measurements were performed by a

single examiner blinded to the group identity. For synaptic

presence, cortical sections immunostained with synaptophysin and

ChAT were examined with densitometry using Image-Pro Plus

image analysis (Media Cybemetics). A total of 25 randomly chosen

areas of cortex, thalamus, and hippocampus from 10 animals per

cohort group were examined at 400Â.

Electrophoretic mobility shift assays

For evaluation of CREB binding to DNA, brain tissue was obtained

and placed in Totex buffer (20 mM HEPES pH 7.9, 350 mM NaCl,

20% glycerol, 1% igepal, 1mM MgCl2, 0.5mM EDTA, 0.1mM

EGTA, 0.1 mM PMSF, 5 mg/ml aprotinin, 50 mM DTT), followed by

cell lysis on ice for 30 min, centrifugation at 14 000 r.p.m. for 15 min

at 4 C, with supernatant retained. Protein levels were determined by

the Bradford method (Biorad) and samples stored at À80 C. Equal

amounts of protein were incubated in a 20-ml reaction mixture

containing 20 mg of bovine serum albumin; 1 mg of poly (dI–dC);

2 ml of buffer containing 20% glycerol, 100 mM KCl, 0.5 mM EDTA,

0.25% NP-40, 2mM dithiothreitol, 0.1% phenylmethylsulphonyl

fluoride and 20mM HEPES, at pH 7.9; 4 ml of buffer containing

20% Ficoll 400, 300mM KCl, 10mM dithiothreitol, 0.1%

phenylmethylsulphonyl fluoride and 100mM HEPES, pH 7.9; and

20000–50000c.p.m. of 32P-labelled oligonucleotide (S) corre-

sponding to a CREB-binding site (50-CAA TGA CAT GCG GCT

ACG TCA CGG CGC AGT GCC C-30). After 20min at room

temperature, reaction products were separated on a 12% non-

denaturing polyacrylamide gel. Radioactivity of dried gels was

detected by exposure to Kodak X-Omat film, and images on the

developed film were scanned into a computer using a UMAX 1200s

scanner. Densitometry was performed using Scion Image software

(Scion Corp., Frederick, MD).

For evaluation of NFkB binding to DNA, nuclear proteins were

extracted with NE-PERTM Nuclear and Cytoplasmic Extraction

Reagents (Pierce, Rockford, IL). Protein–DNA complexes were

detected using biotin end-labelled double-stranded DNA probes

prepared with the Biotin 30 End DNA Labeling Kit (Pierce). The

binding probe used was 50-TCGACAGA[GGGACTTTCC]GAGA

GGC-30, with the binding site indicated in square brackets, and

bold letters indicating regions of variable nucleotides. Electro-

phoretic mobility shift assay (EMSA) was performed with a

LightShift Chemiluminescent EMSA Kit (Pierce). Briefly, nuclear

extracts (10 mg protein) and the 10 Â binding buffer with 2.5%

glycerol, 5 mM MgCl2, 50ng/ml poly(dI-dC), 0.05% NP-40, 1mM

DTT and 20fmol biotin 30-end labelled double-stranded oligonu-

cleotide were incubated at room temperature for 1h in a volume

of 20 ml. For NFkB supershift analysis, an anti–NFkBp65 poly-

clonal antibody (Santa Cruz; 1 mg per reaction) was incubated

with the nuclear proteins on ice for 1h before labelled oligo-

nucleotide was added. Reaction products were separated by

electrophoresis [5% acrylamide (29:1 acryl/bis)] in 0.5 Â TBE.

After electrophoresis, the protein–DNA complexes were trans-

ferred onto nylon membranes and detected using chemilumines-

cence (LightShift kit; Pierce).

Analysis

All statistical comparisons were intended between the following

groups: D I-I and D S-I; D I-I and D I-S; D I-I and C I-I; D S-I

and D S-S; D S-I and C S-I; C I-I and C S-I; C I-I and C I-S; and

C S-I and C S-S. Comparison testing was not performed between

other grouped cohorts, with Bonferroni corrections applied as

appropriate for the above group comparisons.

Data collected in the groups were expressed as mean Æ standard

error. One-way matched/unmatched ANOVA and Student’s t-tests

were performed to compare means between diabetic and control

groups. The Repeated Measures ANOVA assessment was per-

formed for data obtained during the four cognitive studies, as

individual scoring during 1 week partially depended upon

3316

Brain (2008), 131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

performance of the prior week, with the D I-I group compared to

the D S-I and D I-S groups, and the D S-I group compared to the

D I-S and D S-S groups. Also, Area Under The Curve statistical

testing was performed for cognitive testing other than object

recognition tasks. Again, only the groups intended to have

statistical comparisons were analysed as such. For the purposes

of molecular studies and comparisons, one control (non-diabetic)

group was used as a control value, with subsequent comparisons

to other diabetic groups for the molecular test studied.

Results

Intranasal insulin delivery quantification

Quantification of radiolabelled insulin delivery identified

peak delivery to the brain within 1h after I-I delivery and

around 6h with S-I. Systemic insulin concentrations with

intranasal delivery were limited as compared with sub-

cutaneous delivery (Fig. 1). Mice receiving I-I treatment did

not suffer adverse effects throughout the 1-, 2- and 6-h

monitoring periods before sacrifice. S-I delivery led to much

higher concentrations of whole blood insulin and develop-

ment of hypoglycaemia-induced illness in approximately

one-third of mice. Hepatic and kidney insulin levels were

overall two to three times higher in mice administered S-I.

Diabetes

After STZ injection, mice developed diabetes within 2 weeks

in 262/304 (86%) of animals. Diabetic mice were smaller

than non-diabetic mice throughout life beginning 1 month

after STZ injection (Table 1), with D I-I mice maintaining

weight better than D I-S mice. Hyperglycaemia was

identical in D I-I or D I-S mice, but D S-I mice had less

hyperglycaemia and developed hypoglycaemia associated

with mortality in some instances. Non-diabetic (C) S-I mice

also had increased mortality levels relative to C S-S or C I-I

mice (Table 1). Mouse glycated haemoglobin levels were

increased in all diabetic mice at 9 months of life, and were

identical between I-I and I-S mice, but reduced in S-I mice.

The mortality rate in diabetic mice was significantly higher

than in non-diabetic mice, although mice receiving I-I

had improved mortality relative to I-S, S-S and S-I mice

(Table 1) (Kaplan–Meier survival statistics).

The second cohort groups used to complete data within

each intervention group consisted of four mice in each of

the C S-I, D S-S, D S-I, D I-S and D I-I groups. This led to

a data with a minimum of eight mice in each intervention

group (n =9 D I-I, n =10 D I-S, n =9 D S-I, n =10 D S-S,

n =8C I-I, n = 11 C S-I, n =8C S-S, n = 8 C I-S).

Cognitive behavioural data

Cognitive testing continued until 33 weeks of diabetes. All

cognitive data was based upon a minimum of eight mice in

each cohort group at all time points. Learning processes for

each of the tasks appeared to be similar between diabetic

and non-diabetic mice over the first several weeks (Fig. 2).

Diabetic mice performed better than non-diabetic mice in

the first weeks of the Radial Arm Test, as demonstrated

previously (Toth et al., 2006), hypothesized to be due to

hyperphagia contributing to greater exploratory behaviour.

In general, diabetic mice demonstrated waning perfor-

mances on each of the behavioural tasks after 7–10 weeks of

diabetes, with impaired performances continuing through-

out the remainder of the 33 weeks of testing. Both D I-I

and D S-I mice performed better than diabetic cohorts

(D I-S and D S-S) in the Morris Water Maze, Holeboard

Test and Radial Arm Test tasks (Fig. 2), although D I-I

mice consistently outperformed D S-I mice, particularly in

the Morris Water-Maze task. In the Morris Water Maze,

diabetic mice also spent less time in the hemisphere of the

hidden platform (target zone) than non-diabetic mice,

although D I-I mice continued to spend time in the target

zone similar to that of the non-diabetic mice (Supplemen-

tary Fig. 1). Mistakes made in either of the Holeboard or

Radial Arm Tests were also magnified with diabetes and

increased with duration of diabetes, although D I-I mice

made similar numbers of mistakes as non-diabetic mice

(Supplementary Fig. 1). Mouse performance in the Object

Recognition task demonstrated novelty-seeking behaviour

in control mice and D I-I mice, but not in other diabetic

cohort groups (Fig. 2, Supplementary Fig. 1). Additional

assessments using the Area Under the Curve for each

intervention cohort revealed statistically different perfor-

mances on cognitive testing as described in the figure

legends. Analysis provided by the Repeated Measures

ANOVA testing revealed early and late time points for

differences in intervention groups, as demonstrated in

Fig. 2 and Supplementary Fig. 1.

Taken together, these behavioural experiments demon-

strated better maintenance of visuospatial, procedural and

objection recognition memory functioning in the diabetes-

exposed brain receiving intranasal insulin as compared to

the diabetic mouse brain not receiving intranasal insulin.

MRI and brain weight data

Volumetric measurements of brain demonstrated diffuse

cerebral atrophy after 5 months of diabetes (Fig. 3), with

protection demonstrated in D I-I mice after 8 months of

diabetes. Although not identified within individual brain

regions at earlier time points, diabetes-associated atrophy was

detected in the sensorimotor cortex, caudate/putamen,

corpus callosum, internal capsule, CA3 portion of hippo-

campus and cerebral peduncle (Fig. 3) after 8 months of

diabetes. In each of these brain regions, D I-I mice had

measurable protection from atrophy (Fig. 3). Measurement of

brain mass showed evidence of brain atrophy after 5 months

of diabetes, with protection against loss of brain mass first

detected in D I-I mice at 8 months of diabetes (Fig. 3).

As found previously (Toth et al., 2006), there were no

changes in MR T1 or perfusion-weighted imaging measure-

ments identified. Although initial changes could be seen after

5 months, evidence of diabetes-associated leukoencephalopathy

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3317

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

was more easily detected after 8 months of diabetes in all

diabetic mice using quantitative T2 map values. Although

D I-I mice were protected from heightened T2 map values

after 8 months, there were still diabetes-mediated WMA

present in numerous regions of brain described as white

matter regions including corpus callosum and internal

capsule, while similar T2 changes could be identified in

grey matter regions, such as cortex and hippocampus (Fig. 3).

These changes were heterogenous, and were not identified

in other brain regions of interest (Appendix 1).

White matter analysis

Quantitative evaluations of histological sections were

performed to determine myelin presence in brain regions

of interest (Appendix 1). Both LFB preparations and MBP

Fig. 1 Radiolabelled insulin detection. After 1 and 6 h of either I-I (open bars) or S-I (closed bars), I-I led to more rapid and elevated

insulin presence in central nervous system structures, including at cortex and deep brain structures, with much less insulin detected in

blood than with S-I (A). At 6 h after delivery (B), intranasal insulin amounts rose in blood, and subcutaneously delivered insulin had

slowly penetrated nervous system structures, albeit at lesser amounts than with earlier arrival of intranasally delivered insulin.

Significant differences were determined by matched t-tests, with asterisk indicating significant difference (P50.05) between the

intranasal and subcutaneous insulin delivery techniques for each tissue (n = 4 mice in each mouse cohort for each time point).

3318

Brain (2008),131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Table 1 Murine weights, fasting glycaemia levels, glycated haemoglobin levels and survival numbers at induction of diabetes and at harvesting at months

1, 3, 5 and 8 of diabetes

Timepoint

Prior to injection of

STZ/carrier

Month 1

Month 3

Month 5

Month 8

Murine weight

(gÆ SEM)

(gÆ SEM)

(gÆ SEM)

(gÆ SEM)

(gÆ SEM)

Non-diabetic S-S mice

25.6Æ3.2 (n = 25)

32.7Æ3.8 (n = 25)

39.7Æ4.0 (n = 24)

43.4 Æ 4.3Ã (n =24)

47.2 Æ4.9Ã (n=23)

Non-diabetic S-I mice

25.8Æ 3.9 (n = 25)

30.4Æ 4.1 (n = 20)

34.6Æ 4.7 (n = 17)

36.2 Æ 5.1Ã (n =15)

37.0 Æ 5.4Ã (n=12)

Non-diabetic I-S mice

25.1Æ3.7 (n = 25)

32.1Æ3.9 (n = 25)

40.1Æ4.2 (n = 25)

44.7Æ4.6 (n = 25)

48.1Æ5.2 (n = 24)

Non-diabetic I-I mice

25.4Æ3.3 (n = 25)

31.9Æ3.5 (n = 24)

36.2Æ4.9 (n = 23)

39.1Æ5.3 (n = 22)

43.4Æ 4.8 (n = 21)

Diabetic S-S mice

25.2Æ 3.3 (n = 40)

26.9Æ 4.2 (n = 34) (3 non-diabetic)

28.4Æ 4.3 (n = 30)

30.6Æ 5.7 (n = 20)

31.5Æ5.8 (n = 16)

Diabetic S-I mice

25.4Æ3.4 (n = 40)

26.4Æ 4.8 (n = 31) (2 non-diabetic)

27.2Æ 3.8 (n = 24)

28.2Æ4.1 (n = 16)

28.8Æ 4.9 (n = 12)

Diabetic I-S mice

25.2Æ 3.4 (n = 40)

26.2Æ4.8 (n = 33) (3 non-diabetic)

28.9Æ 4.2 (n = 30)

30.2Æ4.9 (n = 22)

30.4 Æ 5.2& (n=18)

Diabetic I-I mice

25.6Æ3.5 (n = 40)

27.8Æ 4.0 (n = 36) (3 non-diabetic)

30.9Æ 4.5 (n = 35)

34.8Æ 3.6 (n = 33)

35.6 Æ 4.9& (n =30)

Murine glycaemia and

8 month glycated haemoglobin

(mmol/l)

(mmol/l)

(mmol/l)

(mmol/l)

(mmol/l)

(percentage of haemoglobin)

Non-diabetic S-S mice

5.5Æ2.3

5.9Æ 2.6

6.1Æ2.7

6.2Æ 3.0

6.6 Æ3.2Ã (12.4%Æ 4.8%)

Non-diabetic S-I mice

5.4Æ 2.6

3.5Æ2.7

3.9Æ 2.9

4.1Æ3.0

4.0 Æ3.1Ã d (9.2%Æ 4.1%)

Non-diabetic I-S mice

6.0Æ2.8

5.9Æ 2.6

5.9Æ 2.8

6.1Æ3.1

6.3Æ2.7 (12.1% Æ 4.7%)

Non-diabetic I-I mice

5.8Æ 2.6

5.7Æ 2.9

5.6Æ3.0

5.7Æ3.0

5.7Æ3.2 (12.6% Æ 4.9%)

Diabetic S-S mice

5.7Æ 2.7

31.7Æ4.9

32.3Æ 6.1

32.2Æ 6.2

32.4 Æ 6.0 (31.6%Æ 6.2%)&

Diabetic S-I mice

5.6Æ2.6

24.7Æ5.2

25.9Æ5.8

24.3Æ5.6

24.8 Æ 6.1 (24.1%Æ 6.6%)&

Diabetic I-S mice

6.1Æ2.9

32.2Æ4.6

32.1Æ5.3

32.1Æ5.2

32.3Æ5.8 (32.0% Æ 6.0%)

Diabetic I-I mice

5.8Æ 2.8

31.5Æ 4.5

31.6Æ 5.0

31.6Æ 5.2

32.0Æ5.6 (30.2% Æ 6.4%)

Murine survival numbers

Number of mice

Number of mice

Number of mice

Number of mice

Number of mice

Non-diabetic S-S mice

25/25 (100%)

25/25 (100%)

24/25 (960%)

24/25 (96%)

23/25 (92%)Ã

Non-diabetic S-I mice

25/25 (100%)

20/25 (80%)

17/25 (68%)

15/25 (60%)

12/25 (48%)Ã d

Non-diabetic I-S mice

25/25 (100%)

25/25 (100%)

25/25 (100%)

25/25 (100%)

24/25 (96%)

Non-diabetic I-I mice

25/25 (100%)

24/25 (96%)

23/25 (92%)

22/25 (88%)

21/25 (84%)

Diabetic S-S mice

40 (100%)

34/37 (3 non-diabetic, 92%)

30/37 (81%)

20/37 (54%)

16/37 (43%)

Diabetic S-I mice

40 (100%)

31/38 (2 non-diabetic, 82%)

24/38 (63%)

16/38 (42%)

12/38 (32%)

Diabetic I-S mice

40 (100%)

33/36 (3 non-diabetic, 92%)

30/36 (83%)

22/36 (61%)

18/36 (50%)&

Diabetic I-I mice

40 (100%)

36/39 (3 non-diabetic, 92%)

35/39 (90%)

33/39 (85%)

30/39 (77%)& f

Glycated haemoglobin values are presented in italics in the 8 month column for glycemia levels. For murine survival, Kaplan^Meier statistics were performed between cohort groups.

All measures are meanÆ SEM for weights, glycemia levels and glycated haemoglobin levels. ÃIndicates significance at P50.05 with comparison of non-diabetic S-S and S-I mice cohort

groups; &Indicates significance between diabetic cohort groups receiving I-S and I-I; ÈIndicates significance with comparison of diabetic I-I mice to S-S and I-S diabetic cohort groups;

dIndicates significance with comparison of non-diabetic S-I mice to all other non-diabetic mice using multiple ANOVA testing with Bonferroni post hoc t-test comparisons (a = 0.016).

Bold values indicate statistical significance as indicated by the superscripted symbols.

In

tran

asal

in

su

lin

an

d

diabetic

en

cep

h

alo

p

athy

Brain

(20

0

8),131,

3311^333

4

3319

at Pennsylvania State University on February 23, 2013 http://brain.oxfordjournals.org/ Downloaded from

Fig. 2 Cognitivebehaviouraldata.Micewithdiabetesareindicatedwitha‘D’,whilemicewithoutdiabetes(controlmice)areindicatedwitha‘C’.

Delivery of subcutaneous saline is indicated as ‘S-S’, subcutaneous insulin as ‘S-I’, intranasal saline as ‘I-S’ and intranasal insulin as ‘I-I’. No baseline

differences existed between any of the mouse cohorts. Morris Water-Maze testing demonstrated learning ability in each cohort, regardless of

diabetes presence (A).Times to reach the platform continued to improve after 7^9 weeks in non-diabetic mice and D I-I mice, whereas the

performance of other diabetic mice failed to improve with prolongation of their later times to complete testing. D S-I mice performed better

than D S-S or D I-S mice at later time points, while D I-I mice outperformed all other diabetic mice after 9 weeks of time. In the Holeboard test,

both D I-I and D S-I mice outperformed other diabetic mouse cohorts after 7 weeks of testing (B). In the initial stages of the Radial ArmTest (C),

3320

Brain (2008),131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Fig. 2 Continued

diabetic mice outperformed non-diabetic mice, perhaps due to enhanced search behaviour related to hyperphagia. After 5 weeks, D I-I and

D S-I mice found their targets faster than other diabetic mouse cohorts, whose performance diminished further over time. After 22

weeks, D I-I mouse times were improved relative to D S-I times. The Object RecognitionTask results are presented as the average of each

of 8 consecutive weeks. This task demonstrated less novelty-seeking behaviour in terms of both visits (D) and time spent (E) at novel

objects in theT2 portion of the experiment for D I-S, D S-I and D S-S mouse cohort groups. For A^C, Repeated Measures ANOVA testing

revealed both early and late time points of significance for D S-I and D I-I mouse cohorts, indicated under the graph with red bars (D S-I)

or blue bars (D I-I) for the weeks where significant differences were identified, when compared to D S-S and D I-S mice, or all other

diabetic mouse cohorts respectively. For A^C, additional Area Under The Curve measurements identified significantly improved perfor-

mances for D I-I mice compared to the D S-I and D I-S groups (P50.025 using Bonferroni corrections). For D and E, significant differences

were determined by multiple ANOVA tests, with asterisk indicating significant difference (P50.016 using Bonferroni corrections) between

the D I-I mouse group and other diabetic mouse cohorts (P50.016 using Bonferroni corrections) for the respective time points (n = 8 ^10

mice in each mouse cohort for each time point).

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3321

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Fig. 3 MRI data. The diabetic murine brain demonstrated atrophy and development of white matter changes over time. There were no

differences between cohort groups for volumetric measurements until after 3 months of diabetes, and no difference between cohort

groups for individual brain region volumetric measurements until 5 months of diabetes. Samples of MR T2 images after 8 months of diabetes

are demonstrated using aged-matched non-diabetic mouse brain (top), diabetic mouse brain receiving long-term intranasal insulin (middle)

and diabetic mouse brain receiving intranasal saline (bottom) (A).Volumetric measurements of the entire brain indicated generalized loss of

brain volume over time in diabetes, first demonstrable after 3^5 months of diabetes (B). D I-I mice were protected from cerebral atrophy

when compared to other diabetic mouse cohorts, detectable after 8 months of diabetes. Individual brain regions also showed atrophy after

8 months of diabetes in diabetic mouse cohort groups other than D I-I mice (C). Wet brain mass paralleled MRI volumetric measurements,

again demonstrating cerebral atrophy in diabetic mice after 5 months, with partial protection found in D I-I mice (D). MR T2 map values

were significantly elevated in both white and grey matter regions for diabetic mice, with D I-I mice again protected when compared to

their diabetic cohort groups (E). Significant differences were determined by multiple ANOVA tests, with asterisk (Ã) indicating significant

difference (P50.0125 using Bonferroni corrections) between the indicated mouse group and other non-diabetic mouse cohorts, while o

indicates a significant difference (P50.0125 using Bonferroni corrections) between the D I-I mouse group and the D I-S and D S-S

cohort groups for the respective time points. Finally, g indicates a significant difference (P50.0125 using Bonferroni corrections)

between the D I-I mouse group and all other diabetic mouse cohort groups for the respective time points (n = 4-6 mice in each

mouse cohort for each time point).

3322

Brain (2008),131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

immunohistochemistry detected reductions in myelin quan-

tity within a number of brain regions of interest (Fig. 4).

As determined previously (Toth et al., 2006), none of the

identified WMA had a pattern which could be attributed to

large vessel infarction. Lost MBP expression closely resembled

abnormalities also identified with LFB staining and with T2

measurement changes found with MR imaging (Fig. 4).

As identified previously (Toth et al., 2006), there were

no differences in neuronal density over grey matter regions

of interest after detailed stereological counts. For example,

neuronal densities within the CA1 region of hippocampus

were 1.73Â 106/mm3

in diabetic mice as compared to

1.79Â 106/mm3 in control mice (P= NS). Neuronal densities

within other grey matter brain regions (Appendix 1) were

similar between diabetic and non-diabetic mouse cohort

groups. Oligodendrocyte counts were performed by exam-

ination of immunohistochemistry for PDGFRa within both

regions of white and grey matter (Appendix 1) and revealed a

44% loss of oligodendrocytes within the internal capsule and

corpus callosum regions in D I-S brains as compared to C I-S

brains, with oligodendrocyte loss protected in D I-I brains,

only demonstrating 17% loss (ANOVA, P50.01).

Fig. 3 Continued.

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3323

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

mRNA and protein quantification

A relative loss of intraneuronal pAkt and decreased nuclear

pAkt presence (Fig. 5) was associated with diabetic enceph-

alopathy. Insulin provision, however, led to amelioration of

this loss in D S-I mice, and particularly in D I-I mice.

Amongst non-diabetic mouse cohorts, there was no

significant difference in pAkt presence in either cortical or

hippocampal neurons. Quantification of Akt and PI3K

mRNA demonstrated downregulation with diabetes in

general, while a near return to normal for both PI3K and

Akt mRNA levels occurred with I-I delivery in diabetic mice

Fig. 4 Quantification of myelin for both Luxol Fast Blue (LFB) and Myelin Basic Protein (MBP). Data for LFB quantification is demonstrated

as (255 ^ Luminosity), as higher luminosity would indicate greater light passage and therefore, less myelin content. LFB loss occurred in

both white and grey matter regions (A) of the diabetic murine brain as compared to the non-diabetic brain (C S-S) (B), with partial

protection present in the D I-I brain (C) when compared to the D I-S brain (D). (Bars = 20mm) MBP loss also occurred in both white and

grey matter regions (E) of the diabetic murine brain as compared to the non-diabetic murine brain (C S-S) (F), with partial protection

again present in the D I-I brain (G) when compared to the D I-S brain (H). (Bars = 10mm) Significant differences were determined by

multiple ANOVA tests, with asterisk (Ã) indicating significant difference (P50.0125 using Bonferroni corrections) between the indicated

mouse group and other non-diabetic mouse cohorts, while g indicates a significant difference (P50.0125 using Bonferroni corrections)

between the D I-I mouse group and all other diabetic mouse cohort groups for the respective time points (n = 4 mice in each mouse

cohort for each time point).

3324

Brain (2008), 131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

(Fig. 5) in cortex (also seen in hippocampus). Protein

quantification in the hippocampus and cortex of diabetic

mice revealed generalized suppression with diabetes, with

partial protection against loss of PI3K, pAkt, GSK3b,

pGSK3b and pCREB (no significant difference for Akt and

CREB) identified in D I-I mice (Fig. 5). Ratios of

phosphorylated to non-phosphorylated downstream mark-

ers revealed increased activation (phosphorylation) of Akt

Fig. 5 The PI3K/Akt pathway is disturbed in the diabetic murine brain. Levels of pAkt fell within hippocampal neurons exposed to dia-

betes, with partial protection provided in D I-I mice (A). Neuronal activation (presence of pAkt in neuronal nuclei) was also diminished in

diabetic hippocampal neurons (B), with intranasal insulin preventing some of this effect (B). Hippocampal neurons in C S-I mice (C)

demonstrated higher levels of nuclear pAkt presence when compared to D I-I mice (D), which were protected when compared to D S-I

mice (E). (Bars = 25mm) Both PI3K and Akt mRNA fell in diabetic murine brain tissue from both hippocampus (data not shown) and

cortex, (F) although D I-I mice were protected against mRNA loss. Similarly, PI3K and Akt protein levels (representative blots,G) fell

based upon semi-quantitative determination of protein levels in both hippocampus and cortex (H). In addition to changes in PI3K and Akt

protein loss, similar protein loss is seen for PI3K/Akt pathway proteins including pAkt, GSK3b, pGSK3b, CREB and pCREB (G,H). There

was no difference in measurements between non-diabetic cohort mouse groups. Quantitative assessment of electrophoretic mobility shift

assays (EMSA) (example in I) demonstated a loss of CREB DNA binding in diabetic hippocampus, reversed with intranasal insulin provision

(J). Meanwhile, phosphorylation ratios for Akt, CREB, and GSK3b within hippocampus were increased with I-I provision in diabetic mice

(K). Quantitative analysis in each situation was performed using three to five samples for each group with multiple ANOVA tests, appro-

priate Bonferroni corrections, and with asterisk indicating significant difference (P50.016) between groups indicated by horizontal bars.

Intranasal insulin and diabetic encephalopathy

Brain (2008), 131, 3311^3334

3325

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

and CREB, along with increased phosphorylation (inactiva-

tion) of GSK3b in the hippocampus for diabetic mice

receiving I-I (Fig. 5). EMSAs demonstrated maintained

DNA binding for CREB in D I-I mice as compared to D I-S

mice (Fig. 5).

Protein and mRNA for portions of the IR, IRb and IR

subsrate (IRS)-1, were also downregulated in the diabetic

murine brain except when I-I was provided (Fig. 6). Both

IRb and IRS-1 were expressed over the neurolemma of

cortical and hippocampal and cortical neurons, with

decreased levels of expression in diabetes, except in D I-I

mice (Fig. 6). For non-diabetic mice, intranasal insulin

provision also led to elevated IRb levels (Fig. 6).

Important components of the synaptic complex, the

vesicular protein synaptophysin (SYP) and the enzyme

choline acetyltransferase (ChAT), were both downregulated

in diabetic hippocampus and cortex (Fig. 7). Again, D I-I

mice were partially spared from loss of synaptic compo-

nents within the cerebrum based upon immunofluorescence

quantification, protein blotting and qRT-PCR studies

(Fig. 7). There were no differences between non-diabetic

cohort groups despite the provision of I-I or S-I (Fig. 7),

and no loss of synaptic components was demonstrated in

the thalamus in any cohort.

Diabetes led to greater accumulation of NFkB and its

greater nuclear presence (activation) (Fig. 8). Both neurons

Fig. 5 Continued.

3326

Brain (2008),131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

and oligodendrocytes in cortical, hippocampal and sub-

cortical regions demonstrated NFkB activation, with

suppressed levels of activation demonstrated in D I-I

mice. As well, D I-I mice also demonstrated partial sup-

pression of overall NFkB mRNA and protein elevation

occurring in both cortex and hippocampus, which exhibited

age-dependent increases over time for both diabetic and

non-diabetic mice (Fig. 8), but greatest in diabetic mice.

Finally, EMSAs demonstrated depressed levels of DNA

binding for NFkB in D I-I mice as compared to D I-S mice,

but levels remained higher in D I-I mice when compared to

non-diabetic mice (Fig. 8). There were no differences in

NFkB levels between non-diabetic cohorts.

Discussion

Intranasal insulin prevented diabetes-mediated cerebral

neurodegeneration without leading to prominent systemic

effects or modification of glycaemia levels. Replacement of

insulin reversed the downregulated PI3K/Akt pathway to

slow or diminish the development of brain atrophy, WMA

and cognitive decline.

Systemic and neural effects of subcutaneous

and intranasal insulin

Insulin delivered through an intranasal route led to

improvements in behavioural (Fig. 2), morphological (Figs

3 and 4), and molecular abnormalities (Figs 5–8) within the

diabetes-exposed brain. I-I did not confer the risks of

systemic hypoglycaemia identified with S-I delivery in this

long-term experimental model. Not only did S-I delivery in

either of diabetic or non-diabetic mice lead to greater

mortality (Table 1) and higher systemic delivery than I-I

(Fig. 1), but failed to improve a number of morphological

and molecular deficits identified with diabetic encepha-

lopathy. While subcutaneous delivery of insulin led to

improved glycated haemoglobin levels at sacrifice, intranasal

insulin delivery did not, suggesting that the beneficial

effects of I-I in diabetic mice was not due to effects upon

hyperglycaemia, which occurred in the D S-I cohort

(Table 1). Whereas systemic insulin enters the brain via a

receptor-mediated, saturable form of transport (Woods

et al., 2003), I-I directly enters the brain bypassing the

blood–brain barrier and travelling by extra-cellular bulk

flow transport along olfactory and trigeminal perivascular

Fig. 6 Diminution of the insulin pathway in the diabetic murine

brain. Levels of IRb fell within the diabetic murine brain but

replenished with I-I, but not S-I (A). mRNA for both IRb and IRS-1

was also deficient in diabetic murine brain regions (hippocampus

not shown) such as cortex (B). Non-diabetic mice receiving

intranasal insulin demonstrated heightened IRb mRNA, and D I-I

mice received protection against loss of both IRb and IRS-1.

Protein levels (C) for IRb and IRS-1 fell in combination with

diabetes as demonstrated with quantitative measurements using

three protein samples for each group (D), with protection again

offered by I-I in both the diabetic cortex and hippocampus.

There were no differences in measurements between non-diabetic

cohort mouse groups. Multiple ANOVA tests, with asterisk indi-

cating significant difference (P50.016) between groups indicated by

horizontal bars, were performed.

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3327

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Fig. 7 Central synaptic components in the diabetic murine brain are decreased. Choline acetyltransferase (ChAT) and synaptophysin (SYP)

levels are diminished within the cortex and hippocampus (data not shown) of diabetic mouse brains relative to non-diabetic murine brains

(A, D). ChAT and SYP levels within the D I-I brain (B, E) were not different from that of the C I-S brain (A, D), while the D I-S brain

demonstrated loss of ChATand SYP (C, F). Losses in synaptic markers were determined by quantification of density for immunostaining for

both ChAT (G) and SYP (H) within cortex and hippocampus (data not shown). qRT-PCR also demonstrated loss of mRNA for both SYP

and ChAT (I). Finally, quantitative assessment of three protein blots (representative blot, J) also portrays a loss of both SYP and ChAT in

the hippocampus and cortex of the diabetic murine brain, where protection against synaptic loss is again provided by I-I to the diabetic

murine brain (K). Multiple ANOVA tests were performed for each comparison, with asterisk indicating significant difference (P50.05)

between groups indicated by horizontal bars.

3328

Brain (2008), 131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Fig. 8 Quantification of NFkB expression. Upregulation in diabetic murine brain for NFkB protein and mRNA occurs as compared to non-

diabetic brain. In the hippocampus, neurons identified with NeuN and oligodendrocytes, identified with PDGFRa, co-express NFkBp65

least in the C S-S brain (A) as compared to diabetic murine brain. D I-I mouse brain (B) had less nuclear activation (nuclear presence) than

D I-S mouse brain (C), within both neurons (D) and oligodendroglia (E). mRNA measurements of NFkBp65 expression identified age-

related increases occurring greatest after 3^5 months of diabetes (F), and again ameliorated with I-I. Protein blotting (representative blot,

G) and its quantitative assessment (H) also demonstrated accumulation of NFkBp65 protein over time, accelerated with diabetes, and

partially suppressed with I-I. The amount of NFkB binding to DNA was significantly upregulated with diabetes (I), with protection from its

increase identied in the D I-I cohort, based upon bands obtained from EMSA, an example of which is demonstrated (J). Quantitative

analysis for Western blots was performed using three samples for each group. For comparisons, multiple ANOVA tests, with asterisk

indicating significant difference between groups indicated by horizontal bars, were performed. All diabetic values were significantly less

than non-diabetic values (significance not visually demonstrated).

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3329

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

channels, as well as axonal transport pathways (Benedict

et al., 2004; Thorne et al., 2004), leading to faster uptake

and less systemic insulin presence.

The role of insulin as a neuroprotective

trophic factor

Insulin, a highly conserved peptide that is no longer

thought of as solely a promoter of glucose turnover, has

now emerged as a key neurotrophic factor in the nervous

system alongside and interacting with the IGF-I receptor

system, also important in maintaining cognitive function

(Trejo et al., 2008). Insulin is a potent trophic factor which

becomes lost within type I diabetes. The major site of

insulin’s activity, IR, is found in high concentration in the

brain, particularly in the cerebral cortex, olfactory bulb,

hippocampus (Fig. 6), amygdala and septum (Havrankova

et al., 1978a, b; Baskin et al., 1987; Unger et al., 1991).

Reduced levels of insulin and its signalling molecules occur

in the CSF and brain of Alzheimer disease patients (Craft

et al., 1998; Hoyer, 2004). Also, IRs are present at synapses

for both astrocytes and neurons (Abbott et al., 1999). As

demonstrated in our study, diabetic rodents also demon-

strate a loss of insulin transduction machinery (Fig. 6),

which has previously been linked to the increased expres-

sion of amyloid precursor protein (APP), APP’s cleavage

enzyme b-secretase, and other abnormalities typical of the

Alzheimer disease brain (Ho et al., 2004; Salkovic-Petrisic

et al., 2006; Li et al., 2007). It has also been demonstrated

that peripheral hyperinsulinaemia and reduced insulin

signalling increase levels of amyloid (Ab) and tau

hyperphosphorylation, leading to the formation of both

senile plaques and neurofibrillary tangles (Zhao et al.,

2004b; Freude et al., 2005). In addition, insulin promotes

physiologic processes critical for memory, including long-

term potentiation, expression of glutamate receptors, and

modulation of neurotransmitter levels (Craft and Watson,

2004). Finally, insulin diminishes hypothalamic–pituitary–

adrenal-axis activity (Hallschmid et al., 2008), recently

speculated to contribute to diabetes-mediated cognitive

dysfunction (Stranahan et al., 2008).

Potentially preventable changes

in diabetic encephalopathy

Our experiments have demonstrated that long-term intra-

nasal insulin delivery protected against brain atrophy

(Fig. 3). Although other studies have reported evidence of

cerebral neuronal loss in diabetes (Li et al., 2007), we could

not detect any loss of neuronal density in this experimen-

tal diabetic brain model, similar to other recent studies

(Stranahan et al., 2008), but we did determine oligoden-

droglial loss occurs and likely relates to the development

of WMA. One potential cause for our detected diabetes-

associated brain atrophy (Fig. 3) is the presence of large

degree of synaptic loss (Fig. 7), also detected in type II

diabetic rat models (Li et al., 2007). Synaptic loss may

precede other forms of pathology in some models of

Alzheimer disease (Yoshiyama et al., 2007), including neu-

ronal loss. Diabetes also leads to a synergistic potentiation

of synaptic loss in transgenic models of Alzheimer disease

(Burdo et al., 2008). This may certainly be impacted by the

presence of insulin receptor at central synapses (Heidenreich

et al., 1983; Matsumoto and Rhoads, 1990; Wan et al., 1997;

Zhao et al., 1999). Signal transduction by neuronal IRs is

exquisitely sensitive to soluble Ab oligomer-mediated disrup-

tion in vitro, and neuronal response to insulin is also inhib-

ited (Zhao et al., 2008). Such synaptic loss identified in our

experiments of the diabetic brain may herald subsequent

neuronal loss, but other models need to be investigated, and

longer duration studies may be necessary to conclude this.

Insulin’s downstream signalling pathways

Insulin is critical for maintenance of numerous downstream

intracellular signalling pathways. Insulin stimulation upre-

gulates protein–tyrosine phosphorylation (Mahadev et al.,

2004) through downstream activation of IRS-2 (Huang

et al., 2005a). Insulin presence leads to activation of Akt

and phosphorylation of Akt substrates (Fig. 5) (Bruss et al.,

2005). In addition, insulin modulates the inner mitochon-

drial membrane potential through activation of the PI3K

pathway, stimulating phosphorylation of Akt and cAMP

response element-binding protein CREB (Marshall, 1995;

Yao and Cooper, 1995; Fernyhough et al., 2003; Viard et al.,

2004; Huang et al., 2005b), as well as supporting neuritic

extension and branching (Jones et al., 2003). PI3K also

enhances voltage-dependent calcium channel current func-

tioning in diabetic neurons (Viard et al., 2004). In our

studies, downregulation of PI3K/Akt occurred throughout

the diabetic murine brain (Fig. 5), while insulin delivered

intranasally, and to a lesser extent, subcutaneously,

prevented such downregulation while concurrently leading

to improvements in behaviour and morphology.

Insulin’s benefits in the diabetic brain may relate to

inactivation of GSK-3b and activation of CREB. Besides

regulating the transcriptional activities of CREB (Cohen

and Frame, 2001; Grimes and Jope, 2001), GSK-3b is also a

neuron-specific (Leroy and Brion, 1999) apoptosis pro-

moter in the non-phosphorylated, or active state (Hetman

et al., 2000); Akt-mediated phosphorylation of GSK-3b

renders it inactive, giving anti-apoptotic properties (Pap

and Cooper, 1998; Bijur et al., 2000; Hetman et al., 2000).

Similar to insulin in our study, IGF-I also activates the

PI3K/Akt pathway (Dudek et al., 1997; Zheng et al., 2002),

leading to phosphorylation of CREB and GSK-3b

(Leinninger et al., 2004). IGF-I provision also induces oli-

godendrocyte progenitor proliferation via Akt activation

(Cui and Almazan, 2007), with GSK-3b phosphorylation in

oligodendrocyte progenitor cells affecting oligodendrocyte

stability (Frederick et al., 2007). pGSK-3b may also regulate

gene expression and activity of transcriptional factor

3330

Brain (2008),131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

binding to the MAG promoter region (Ogata et al., 2004),

promoting myelination and possibly explaining insulin-

mediated protection against WMA development in the

experimental diabetic murine brain (Figs 3 and 4). While

CREB phosphorylation inhibits apoptosis in neurons

(Walton et al., 1999), the loss of CREB results in impaired

axonal growth (Lonze et al., 2002), suggesting that CREB is

neuroprotective. Intranasal insulin led to heightened

pCREB levels as well as greater CREB DNA binding

indicating transcription (Fig. 5). Overall, we hypothesize

that insulin’s benefits in the diabetic murine brain are due

to maintained phosphorylation of CREB and GSK-3b.

The utility of intranasal delivery in

diabetic encephalopathy

In the case of diabetic encephalopathy, and likely the

Alzheimer disease brain, insulin’s trophic effects are lacking.

I-I delivery without modifying systemic glucose or glycated

haemoglobin (Table 1) (Tomlinson and Gardiner, 2008)

appears to be essential to insulin’s long-term benefits in

brain exposed to diabetes. Human studies have already

demonstrated safe administration of intranasal insulin in

patients with Alzheimer disease, leading to some improve-

ment in cognition and modulating markers of Alzheimer

disease (Reger et al., 2006, 2008). These studies have

concentrated upon the potential benefits of intranasal

insulin delivery upon clinical and pathological markers of

Alzheimer disease. Other clinical studies examining intra-

nasal insulin delivery have also demonstrated effects upon

the CNS, such as improvement in memory (Benedict et al.,

2007, 2008), lowering of food intake (Tomlinson et al.,

2008), and improvement in mood (Hallschmid et al., 2008).

In humans, plasma glucose levels may be minimally

impacted by intranasal insulin delivery (Born et al., 2002;

Reger et al., 2006, 2008; Tomlinson et al., 2008). As of yet,

intranasal insulin delivery to diabetic subjects for the intent

of improving memory or other diabetic complications has

not yet been described, although intranasal insulin delivery

has been examined as a potential method for the treatment

of diabetes itself (Owens et al., 2003; Khafagy et al., 2007).

Our study results must be considered under the

limitations of working in a murine model, and the inability

to achieve a long-term model of murine type I diabetes

with optimal glycaemic management as a suitable control

group. The mouse cohorts were subjected to intensive

testing throughout their lifetime, with the possibility of

stressful impacts upon the results obtained. The use of

multiple cognitive studies, often needed as part of a battery

of tests, may have led to crossover effects affecting results

from one cognitive test based upon the preceding test. It is

also possible that hypoglycaemia may have impacted some

of the results of cognitive testing, and the impact of

hypoglycaemia upon the D I-S and C I-S cohort groups was

anticipated, but difficult to avoid. As well, performance of

behavioural testing within multiple cohorts of mice within

each intervention group may have contributed to perfor-

mance disparities. Based upon our studies, it is difficult to

develop a more appropriate control group of diabetic mice

with long-term glycaemic control based upon the STZ-

induced diabetic model. Despite these difficulties, D S-I

mice performed better than D S-S cohort mice during

portions of cognitive testing. Although cognitive changes in

our mice also seem to occur mainly during the younger and

older age time points (similar to expected changes in type I

diabetic patients), explanations for a relative plateau of

function during the middle-aged years are unknown.

(Wessels et al., 2007, 2008; Biessels et al., 2008;

Kloppenborg et al., 2008;) At this time, our results are

limited to murine models of type I diabetes, and other

forms of diabetes, including models of type II diabetes, will

need study in the future to confirm our results in other

models of diabetes.

Conclusions

Intranasal insulin delivery is a potential therapy to

ameliorate behavioural, morphological and molecular

changes occurring in brain exposed to diabetes over time.

Our results provide strong evidence for benefits of insulin

without impact upon serum glucose levels, indicating that

insulin is an important neurotrophic factor in the manage-

ment of diabetes-mediated brain disease. These data

support the development of clinical studies for the

prevention and slowing of the adverse effects of diabetes

upon the brain using intranasal delivery of insulin.

Supplementary material

Supplementary material is available at Brain online.

Funding

This study was supported by an operating grant from the

Alberta Heritage Foundation for Medical Research and the

Canadian Diabetes Association. C.T. is a Clinical

Investigator of the Alberta Heritage Foundation for

Medical Research and D.W.Z. is a Scientist of the Alberta

Heritage Foundation for Medical Research.

References

Abbott MA, Wells DG, Fallon JR. The insulin receptor tyrosine kinase

substrate p58/53 and the insulin receptor are components of CNS

synapses. J Neurosci 1999; 19: 7300–7308.

Adamo M, Raizada MK, LeRoith D. Insulin and insulin-like growth factor

receptors in the nervous system. Mol Neurobiol 1989; 3: 71–100.

Akisaki T, Sakurai T, Takata T, Umegaki H, Araki A, Mizuno S, et al.

Cognitive dysfunction associates with white matter hyperintensities and

subcortical atrophy on magnetic resonance imaging of the elderly

diabetes mellitus Japanese elderly diabetes intervention trial (J-EDIT).

Diabetes Metab Res Rev 2006; 22: 376–84.

Awad N, Gagnon M, Messier C. The relationship between impaired

glucose tolerance, type 2 diabetes, and cognitive function. J Clin Exp

Neuropsychol 2004; 26: 1044–80.

Intranasal insulin and diabetic encephalopathy

Brain (2008), 131, 3311^3334

3331

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Baskin DG, Figlewicz DP, Woods SC, Porte D Jr, Dorsa DM. Insulin in the

brain. Annu Rev Physiol 1987; 49: 335–47.

Benedict C, Hallschmid M, Hatke A, Schultes B, Fehm HL, Born J, et al.

Intranasal

insulin

improves

memory

in

humans.

Psychoneuroendocrinology 2004; 29: 1326–34.

Benedict C, Hallschmid M, Schultes B, Born J, Kern W. Intranasal insulin

to improve memory function in humans. Neuroendocrinology 2007; 86:

136–42.

Biessels GJ, Deary IJ, Ryan CM. Cognition and diabetes: a lifespan

perspective. Lancet Neurol 2008; 7: 184–90.

Biessels GJ, Kamal A, Ramakers GM, Urban IJ, Spruijt BM, Erkelens DW,

et al. Place learning and hippocampal synaptic plasticity in streptozo-

tocin-induced diabetic rats. Diabetes 1996; 45: 1259–66.

Biessels GJ, Kamal A, Urban IJ, Spruijt BM, Erkelens DW, Gispen WH.

Water maze learning and hippocampal synaptic plasticity in

streptozotocin-diabetic rats: effects of insulin treatment. Brain Res

1998; 800: 125–35.

Bijur GN, De SP, Jope RS. Glycogen synthase kinase-3beta facilitates

staurosporine- and heat shock-induced apoptosis. Protection by lithium.

J Biol Chem 2000; 275: 7583–90.

Born J, Lange T, Kern W, McGregor GP, Bickel U, Fehm HL. Sniffing

neuropeptides: a transnasal approach to the human brain. Nat Neurosci

2002; 5: 514–6.

Bruss MD, Arias EB, Lienhard GE, Cartee GD. Increased phosphorylation

of Akt substrate of 160 kDa (AS160) in rat skeletal muscle in response to

insulin or contractile activity. Diabetes 2005; 54: 41–50.

Burdo JR, Chen Q, Calcutt NA, Schubert D. The pathological interaction

between diabetes and presymptomatic Alzheimer’s disease. Neurobiol

Aging 2008.

Chen XQ, Fawcett JR, Rahman YE, Ala TA, Frey II WH. Delivery of nerve

growth factor to the brain via the olfactory pathway. J Alzheimers Dis

1998; 1: 35–44.

Cohen RA. Dysfunction of vascular endothelium in diabetes mellitus.

Circulation 1993; 87: V67–76.

Cohen P, Frame S. The renaissance of GSK3. Nat Rev Mol Cell Biol 2001;

2: 769–76.

Craft S, Peskind E, Schwartz MW, Schellenberg GD, Raskind M,

Porte D, Jr. Cerebrospinal fluid and plasma insulin levels in

Alzheimer’s disease: relationship to severity of dementia and apolipo-

protein E genotype. Neurology 1998; 50: 164–8.

Craft S, Watson GS. Insulin and neurodegenerative disease: shared and

specific mechanisms. Lancet Neurol 2004; 3: 169–78.

Crawley JN: Behavioral paradigms. The Morris water task. In: Crawley JN,

editor. What’s wrong with my mouse? Behavioral phenotyping of

transgenic and knockout mice. New York: Wiley-Liss; 2000. p. 87–95.

Cui QL, Almazan G. IGF-I-induced oligodendrocyte progenitor prolifera-

tion requires PI3K/Akt, MEK/ERK, and Src-like tyrosine kinases. J

Neurochem 2007; 100: 1480–93.

de Groot JC, de Leeuw FE, Oudkerk M, van GJ, Hofman A, Jolles J,

Breteler MM. Cerebral white matter lesions and cognitive function: the

Rotterdam Scan Study. Ann Neurol 2000; 47: 145–51.

Dhanda DS, Frey WH, Leopold D, Kompella UB. Nose to brain delivery:

approaches for drug deposition in the human olfactory epithelium.

Drug Deliv Technol 2005; 5: 64–72.

Dudek H, Datta SR, Franke TF, Birnbaum MJ, Yao R, Cooper GM, et al.

Regulation of neuronal survival by the serine-threonine protein kinase

Akt. Science 1997; 275: 661–5.

Fernyhough P, Huang TJ, Verkhratsky A. Mechanism of mitochondrial

dysfunction in diabetic sensory neuropathy. J Peripher Nerv Syst 2003;

8: 227–35.

File SE, Wardill AG. The reliability of the hole-board apparatus.

Psychopharmacologia 1975; 44: 47–51.

Fox MA, Chen RS, Holmes CS. Gender differences in memory and

learning in children with insulin-dependent diabetes mellitus (IDDM)

over a 4-year follow-up interval. J Pediatr Psychol 2003; 28: 569–78.

Frederick TJ, Min J, Altieri SC, Mitchell NE, Wood TL. Synergistic

induction of cyclin D1 in oligodendrocyte progenitor cells by IGF-I and

FGF-2 requires differential stimulation of multiple signaling pathways.

Glia 2007; 55: 1011–22.

Freude S, Plum L, Schnitker J, Leeser U, Udelhoven M, Krone W, et al.

Peripheral hyperinsulinemia promotes tau phosphorylation in vivo.

Diabetes 2005; 54: 3343–8.

Grimes CA, Jope RS. CREB DNA binding activity is inhibited by glycogen

synthase kinase-3 beta and facilitated by lithium. J Neurochem 2001; 78:

1219–32.

Gundersen HJ, Bendtsen TF, Korbo L, Marcussen N, Moller A, Nielsen K,

et al. Some new, simple and efficient stereological methods and their use

in pathological research and diagnosis. APMIS 1988; 96: 379–94.

Haan MN. Therapy Insight: type 2 diabetes mellitus and the risk of late-

onset Alzheimer’s disease. Nat Clin Pract Neurol 2006; 2: 159–66.

Hallschmid M, Benedict C, Schultes B, Born J, Kern W. Obese men

respond to cognitive but not to catabolic brain insulin signaling. Int J

Obes (Lond) 2008; 32: 275–82.

Hattori Y, Hattori S, Sato N, Kasai K. High-glucose-induced nuclear factor

kappaB activation in vascular smooth muscle cells. Cardiovasc Res 2000;

46: 188–97.

Havrankova J, Schmechel D, Roth J, Brownstein M. Identification of

insulin in rat brain. Proc Natl Acad Sci USA 1978a; 75: 5737–41.

Havrankova J, Roth J, Brownstein M. Insulin receptors are widely

distributed in the central nervous system of the rat. Nature 1978b;

272: 827–9.

Heidenreich KA, Zahniser NR, Berhanu P, Brandenburg D, Olefsky JM.

Structural differences between insulin receptors in the brain and

peripheral target tissues. J Biol Chem 1983; 258: 8527–30.

Hetman M, Cavanaugh JE, Kimelman D, Xia Z. Role of glycogen synthase

kinase-3beta in neuronal apoptosis induced by trophic withdrawal.

J Neurosci 2000; 20: 2567–74.

Ho L, Qin W, Pompl PN, Xiang Z, Wang J, Zhao Z, et al. Diet-induced

insulin resistance promotes amyloidosis in a transgenic mouse model of

Alzheimer’s disease. FASEB J 2004; 18: 902–4.

Hoyer S. Glucose metabolism and insulin receptor signal transduction in

Alzheimer disease. Eur J Pharmacol 2004; 490: 115–25.

Huang C, Thirone AC, Huang X, Klip A. Differential contribution of

insulin receptor substrates 1 versus 2 to insulin signaling and glucose

uptake in l6 myotubes. J Biol Chem 2005a; 280: 19426–35.

Huang TJ, Verkhratsky A, Fernyhough P. Insulin enhances mitochondrial

inner membrane potential and increases ATP levels through phosphoi-

nositide 3-kinase in adult sensory neurons. Mol Cell Neurosci 2005b; 28:

42–54.

Ikram MA, Vrooman HA, Vernooij MW, van der LF, Hofman A, van

der LA, et al. Brain tissue volumes in the general elderly population. The

Rotterdam Scan Study. Neurobiol Aging 2008; 29: 882–90.

Jacobson AM, Musen G, Ryan CM, Silvers N, Cleary P, Waberski B, et al.

Long-term effect of diabetes and its treatment on cognitive function.

N Engl J Med 2007; 356: 1842–52.

Jones DM, Tucker BA, Rahimtula M, Mearow KM. The synergistic effects

of NGF and IGF-1 on neurite growth in adult sensory neurons:

convergence on the PI 3-kinase signaling pathway. J Neurochem 2003;

86: 1116–28.

Kaiyala KJ, Prigeon RL, Kahn SE, Woods SC, Schwartz MW. Obesity

induced by a high-fat diet is associated with reduced brain insulin

transport in dogs. Diabetes 2000; 49: 1525–33.

Kamal A, Biessels GJ, Duis SE, Gispen WH. Learning and hippocampal

synaptic plasticity in streptozotocin-diabetic rats: interaction of diabetes

and ageing. Diabetologia 2000; 43: 500–6.

Kamal A, Biessels GJ, Gispen WH, Ramakers GM. Synaptic transmission

changes in the pyramidal cells of the hippocampus in streptozotocin-

induced diabetes mellitus in rats. Brain Res 2006; 1073–1074: 276–80.

Khafagy E, Morishita M, Onuki Y, Takayama K. Current challenges in

non-invasive insulin delivery systems: a comparative review. Adv Drug

Deliv Rev 2007; 59: 1521–46.

Kloppenborg RP, van den Berg E, Kappelle LJ, Biessels GJ. Diabetes and

other vascular risk factors for dementia: which factor matters most? A

systematic review. Eur J Pharmacol 2008; 585: 97–108.

3332

Brain (2008), 131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Knopman D, Boland LL, Mosley T, Howard G, Liao D, Szklo M, et al.

Cardiovascular risk factors and cognitive decline in middle-aged adults.

Neurology 2001; 56: 42–8.

Kramer L, Fasching P, Madl C, Schneider B, Damjancic P, Waldhausl W,

et al. Previous episodes of hypoglycemic coma are not associated with

permanent cognitive brain dysfunction in IDDM patients on intensive

insulin treatment. Diabetes 1998; 47: 1909–14.

Last D, Alsop DC, Abduljalil AM, Marquis RP, de BC, Hu K, et al. Global

and regional effects of type 2 diabetes on brain tissue volumes and

cerebral vasoreactivity. Diabetes Care 2007; 30: 1193–9.

Leinninger GM, Backus C, Uhler MD, Lentz SI, Feldman EL. Phosphati-

dylinositol 3-kinase and Akt effectors mediate insulin-like growth factor-I

neuroprotection in dorsal root ganglia neurons. FASEB J 2004; 18: 1544–6.

Leroy K, Brion JP. Developmental expression and localization of

glycogen synthase kinase-3beta in rat brain. J Chem Neuroanat 1999;

16: 279–3.

Li ZG, Zhang W, Sima AA. The role of impaired insulin/IGF action in

primary diabetic encephalopathy. Brain Res 2005; 1037: 12–24.

Li ZG, Zhang W, Sima AA. Alzheimer-like changes in rat models of

spontaneous diabetes. Diabetes 2007; 56: 1817–24.

Liu XF, Fawcett JR, Hanson LR, Frey WH. The window of opportunity for

treatment of focal cerebral ischemic damage with noninvasive intranasal

insulin-like growth factor-I in rats. J Stroke Cerebrovasc Dis 2004; 13:

16–23.

Lonze BE, Riccio A, Cohen S, Ginty DD. Apoptosis, axonal growth defects,

and degeneration of peripheral neurons in mice lacking CREB. Neuron

2002; 34: 371–85.

Lupien SB, Bluhm EJ, Ishii DN. Effect of IGF-I on DNA, RNA, and

protein loss associated with brain atrophy and impaired learning in

diabetic rats. Neurobiol Dis 2006; 21: 487–95.

Mahadev K, Wu X, Motoshima H, Goldstein BJ. Integration of multiple

downstream signals determines the net effect of insulin on MAP kinase

vs. PI 30-kinase activation: potential role of insulin-stimulated H(2)O(2).

Cell Signal 2004; 16: 323–31.

Manschot SM, Brands AM, van der GJ, Kessels RP, Algra A, Kappelle LJ,

et al. Brain magnetic resonance imaging correlates of impaired cognition

in patients with type 2 diabetes. Diabetes 2006; 55: 1106–13.

Marshall CJ. Specificity of receptor tyrosine kinase signaling: transient

versus sustained extracellular signal-regulated kinase activation. Cell

1995; 80: 179–85.

Matsumoto H, Rhoads DE. Specific binding of insulin to membranes from

dendrodendritic synaptosomes of rat olfactory bulb. J Neurochem 1990;

54: 347–50.

McCarthy AM, Lindgren S, Mengeling MA, Tsalikian E, Engvall JC. Effects

of diabetes on learning in children. Pediatrics 2002; 109: E9.

Miles WR, Root HF. Psychologic tests applied to diabetic patients. Arch

Intern Med 1922; 30: 767–77.

Mooradian AD. Central nervous system complications of diabetes

mellitus–a perspective from the blood-brain barrier. Brain Res Brain

Res Rev 1997; 23: 210–8.

Morris R. Developments of a water-maze procedure for studying spatial

learning in the rat. J Neurosci Methods 1984; 11: 47–60.

Munoz DG, Hastak SM, Harper B, Lee D, Hachinski VC. Pathologic

correlates of increased signals of the centrum ovale on magnetic

resonance imaging. Arch Neurol 1993; 50: 492–7.

Musen G, Lyoo IK, Sparks CR, Weinger K, Hwang J, Ryan CM, et al.

Effects of type 1 diabetes on gray matter density as measured by voxel-

based morphometry. Diabetes 2006; 55: 326–33.

Murray AD, Staff RT, Shenkin SD, Deary IJ, Starr JM, Whalley LJ. Brain

white matter hyperintensities: relative importance of vascular risk factors

in nondemented elderly people. Radiology 2005; 237: 251–7.

Northam EA, Anderson PJ, Jacobs R, Hughes M, Warne GL, Werther GA.

Neuropsychological profiles of children with type 1 diabetes 6 years after

disease onset. Diabetes Care 2001; 24: 1541–6.

Ogata T, Iijima S, Hoshikawa S, Miura T, Yamamoto S, Oda H, et al.

Opposing extracellular signal-regulated kinase and Akt pathways control

Schwann cell myelination. J Neurosci 2004; 24: 6724–32.

Owens DR, Zinman B, Bolli G. Alternative routes of insulin delivery.

Diabet Med 2003; 20: 886–98.

Pantoni L, Garcia JH. Pathogenesis of leukoaraiosis: a review. Stroke 1997;

28: 652–9.

Pantoni L, Poggesi A, Inzitari D. The relation between white-matter lesions

and cognition. Curr Opin Neurol 2007; 20: 390–7.

Pap M, Cooper GM. Role of glycogen synthase kinase-3 in the

phosphatidylinositol 3-Kinase/Akt cell survival pathway. J Biol Chem

1998; 273: 19929–32.

Patti ME, Sun XJ, Bruening JC, Araki E, Lipes MA, White MF, et al.

4PS/insulin receptor substrate (IRS)-2 is the alternative substrate of the

insulin receptor in IRS-1-deficient mice. J Biol Chem 1995; 270:

24670–3.

Reaven GM, Thompson LW, Nahum D, Haskins E. Relationship between

hyperglycemia and cognitive function in older NIDDM patients.

Diabetes Care 1990; 13: 16–21.

Reger MA, Watson GS, Frey WH, Baker LD, Cholerton B, Keeling ML,

et al. Effects of intranasal insulin on cognition in memory-impaired

older adults: modulation by APOE genotype. Neurobiol Aging 2006; 27:

451–8.

Reger MA, Watson GS, Green PS, Wilkinson CW, Baker LD, Cholerton B,

et al. Intranasal insulin improves cognition and modulates beta-amyloid

in early AD. Neurology 2008; 70: 440–8.

Ross TM, Martinez PM, Renner JC, Thorne RG, Hanson LR, Frey WH.

Intranasal administration of interferon beta bypasses the blood-brain

barrier to target the central nervous system and cervical lymph nodes: a

non-invasive treatment strategy for multiple sclerosis. J Neuroimmunol

2004; 151: 66–77.

Rothblat LA, Kromer LF. Object recognition memory in the rat: the role of

the hippocampus. Behav Brain Res 1991; 42: 25–32.

Ryan CM. Neurobehavioral complications of type I diabetes. Examination

of possible risk factors. Diabetes Care 1988; 11: 86–93.

Ryan CM, Williams TM. Effects of insulin-dependent diabetes on learning

and memory efficiency in adults. J Clin Exp Neuropsychol 1993; 15:

685–700.

Ryan CM, Williams TM, Finegold DN, Orchard TJ. Cognitive dysfunction

in adults with type 1 (insulin-dependent) diabetes mellitus of long

duration: effects of recurrent hypoglycaemia and other chronic

complications. Diabetologia 1993; 36: 329–34.

Ryan CM, Geckle M. Why is learning and memory dysfunction in Type 2

diabetes limited to older adults? Diabetes Metab Res Rev 2000; 16:

308–15.

Salkovic-Petrisic M, Tribl F, Schmidt M, Hoyer S, Riederer P. Alzheimer-

like changes in protein kinase B and glycogen synthase kinase-3 in rat

frontal cortex and hippocampus after damage to the insulin signalling

pathway. J Neurochem 2006; 96: 1005–15.

Schoenle EJ, Schoenle D, Molinari L, Largo RH. Impaired intellectual

development in children with Type I diabetes: association with HbA(1c),

age at diagnosis and sex. Diabetologia 2002; 45: 108–14.

Schmidt R, Launer LJ, Nilsson LG, Pajak A, Sans S, Berger K, et al. Magnetic

resonance imaging of the brain in diabetes: the Cardiovascular

Determinants of Dementia (CASCADE) Study. Diabetes 2004; 53: 687–92.

Schwabe K, Klein S, Koch M. Behavioural effects of neonatal lesions of the

medial prefrontal cortex and subchronic pubertal treatment with

phencyclidine of adult rats. Behav Brain Res 2006; 168: 150–60.

Schwartz AV, Vittinghoff E, Sellmeyer DE, Feingold KR, de RN,

Strotmeyer ES, et al. Diabetes-related complications, glycemic control,

and falls in older adults. Diabetes Care 2008; 31: 391–6.

Stewart R, Liolitsa D. Type 2 diabetes mellitus, cognitive impairment and

dementia. Diabet Med 1999; 16: 93–112.

Strachan MW, Deary IJ, Ewing FM, Frier BM. Is type II diabetes associated

with an increased risk of cognitive dysfunction? A critical review of

published studies. Diabetes Care 1997; 20: 438–45.

Stranahan AM, Arumugam TV, Cutler RG, Lee K, Egan JM, Mattson MP.

Diabetes impairs hippocampal function through glucocorticoid-

mediated effects on new and mature neurons. Nat Neurosci 2008; 11:

309–17.

Intranasal insulin and diabetic encephalopathy

Brain (2008),131, 3311^3334

3333

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from

Thorne RG, Pronk GJ, Padmanabhan V, Frey WH. Delivery of insulin-like

growth factor-I to the rat brain and spinal cord along olfactory and

trigeminal pathways following intranasal administration. Neuroscience

2004; 127: 481–96.

Tomlinson DR, Gardiner NJ. Glucose neurotoxicity. Nat Rev Neurosci

2008; 9: 36–45.

Benedict C, Kern W, Schultes B, Born J, Hallschmid M. Differential

sensitivity of men and women to anorexigenic and memory improving

effects of intranasal insulin. J Clin Endocrinol Metab 2008.

Toth C, Schmidt AM, Tuor UI, Francis G, Foniok T, Brussee V, et al.

Diabetes, leukoencephalopathy and RAGE. Neurobiol Dis 2006; 23:

445–61.

Trejo JL, Llorens-Martın MV, Torres-Alemán I. The effects of exercise on

spatial learning and anxiety-like behavior are mediated by an IGF-I-

dependent mechanism related to hippocampal neurogenesis. Mol Cell

Neurosci. 2008; 37: 402–11.

Tullberg M, Fletcher E, DeCarli C, Mungas D, Reed BR, Harvey DJ, et al.

White matter lesions impair frontal lobe function regardless of their

location. Neurology 2004; 63: 246–53.

Unger J, McNeill TH, Moxley RT, III, White M, Moss A, Livingston JN.

Distribution of insulin receptor-like immunoreactivity in the rat

forebrain. Neuroscience 1989; 31: 143–57.

Unger JW, Livingston JN, Moss AM. Insulin receptors in the central

nervous system: localization, signalling mechanisms and functional

aspects. Prog Neurobiol 1991; 36: 343–62.

van Harten B, de Leeuw FE, Weinstein HC, Scheltens P, Biessels GJ. Brain

imaging in patients with diabetes: a systematic review. Diabetes Care

2006; 29: 2539–48.

van Harten B, Oosterman JM, Potter van Loon BJ, Scheltens P,

Weinstein HC. Brain lesions on MRI in elderly patients with type 2

diabetes mellitus. Eur Neurol 2007; 57: 70–4.

Verdelho A, Madureira S, Ferro JM, Basile AM, Chabriat H, Erkinjuntti T,

et al. Differential impact of cerebral white matter changes, diabetes,

hypertension and stroke on cognitive performance among non-disabled

elderly. The LADIS study. J Neurol Neurosurg Psychiatry 2007; 78:

1325–30.

Viard P, Butcher AJ, Halet G, Davies A, Nurnberg B, Heblich F,

Dolphin AC. PI3K promotes voltage-dependent calcium channel

trafficking to the plasma membrane. Nat Neurosci 2004; 7: 939–46.

Vig PJ, Subramony SH, D’Souza DR, Wei J, Lopez ME. Intranasal

administration of IGF-I improves behavior and Purkinje cell pathology

in SCA1 mice. Brain Res Bull 2006; 69: 573–9.

Walton M, Woodgate AM, Muravlev A, Xu R, During MJ, Dragunow M.

CREB phosphorylation promotes nerve cell survival. J Neurochem 1999;

73: 1836–42.

Wan Q, Xiong ZG, Man HY, Ackerley CA, Braunton J, Lu WY, et al.

Recruitment of functional GABA(A) receptors to postsynaptic domains

by insulin. Nature 1997; 388: 686–90.

Weinger K, Jacobson AM, Musen G, Lyoo IK, Ryan CM, Jimerson DC,

et al. The effects of type 1 diabetes on cerebral white matter.

Diabetologia 2008; 51: 417–25.

Wessels AM, Rombouts SA, Remijnse PL, Boom Y, Scheltens P, Barkhof F,

et al. Cognitive performance in type 1 diabetes patients is associated

with cerebral white matter volume. Diabetologia 2007; 50: 1763–9.

White MF, Yenush L. The IRS-signaling system: a network of docking

proteins that mediate insulin and cytokine action. Curr Top Microbiol

Immunol 1998; 228: 179–208.

Whitman GT, Tang Y, Lin A, Baloh RW. A prospective study of cerebral

white matter abnormalities in older people with gait dysfunction.

Neurology 2001; 57: 990–4.

Wickelgren I. Tracking insulin to the mind. Science 1998; 280: 517–9.

Woods SC, Seeley RJ, Baskin DG, Schwartz MW. Insulin and the blood-

brain barrier. Curr Pharm Des 2003; 9: 795–800.

Yao R, Cooper GM. Requirement for phosphatidylinositol-3 kinase in the

prevention of apoptosis by nerve growth factor. Science 1995; 267: 2003–6.

Yoshiyama Y, Higuchi M, Zhang B, Huang SM, Iwata N, Saido TC, et al.

Synapse loss and microglial activation precede tangles in a P301S

tauopathy mouse model. Neuron 2007; 53: 337–51.

Yu XR, Jia GR, Gao GD, Wang SH, Han Y, Cao W. Neuroprotection of

insulin against oxidative stress-induced apoptosis in cultured retinal

neurons: involvement of phosphoinositide 3-kinase/Akt signal pathway.

Acta Biochim Biophys Sin (Shanghai) 2006; 38: 241–8.

Zhao W, Chen H, Xu H, Moore E, Meiri N, Quon MJ, et al. Brain insulin

receptors and spatial memory. Correlated changes in gene expression,

tyrosine phosphorylation, and signaling molecules in the hippocampus

of water maze trained rats. J Biol Chem 1999; 274: 34893–902.

Zhao WQ, Chen H, Quon MJ, Alkon DL. Insulin and the insulin receptor

in experimental models of learning and memory. Eur J Pharmacol

2004a; 490: 71–81.

Zhao L, Teter B, Morihara T, Lim GP, Ambegaokar SS, Ubeda OJ, et al.

Insulin-degrading enzyme as a downstream target of insulin receptor

signaling cascade: implications for Alzheimer’s disease intervention.

J Neurosci 2004b; 24: 11120–6.

Zhao WQ, De Felice FG, Fernandez S, Chen H, Lambert MP, Quon MJ,

et al. Amyloid beta oligomers induce impairment of neuronal insulin

receptors. FASEB J 2008; 22: 246–60.

Zheng WH, Kar S, Quirion R. Insulin-like growth factor-1-induced

phosphorylation of transcription factor FKHRL1 is mediated by

phosphatidylinositol 3-kinase/Akt kinase and role of this pathway in

insulin-like growth factor-1-induced survival of cultured hippocampal

neurons. Mol Pharmacol 2002; 62: 225–33.

Appendix 1

The following brain regions were chosen as areas for close

inspection within this study for MRI and myelin quanti-

fication portions of the experiment:

Caudate putamen

Primary motor/sensory cortex

Internal capsule

Cerebral peduncle

CA1 region of hippocampus

CA2/3 region of hippocampus

Ventroposterior region of thalamus

Corpus callosum

Amygdala

Substantia nigra pars reticulata

Subiculum

Lentiform nuclear region

Primary visual cortex

Cerebellum

3334

Brain (2008),131, 3311^3334

G. J. Francis et al.

at Pennsylvania State University on February 23, 2013

http://brain.oxfordjournals.org/

Downloaded from