NeuroImage 56 (2011) 1286–1292

Contents lists available at ScienceDirect

NeuroImage

j ourna l homepage: www.e lsev ie r.com/ locate /yn img

Quantitative in vivo measurement of early axonal transport deficits in a tripletransgenic mouse model of Alzheimer's disease using manganese-enhanced MRI

Jieun Kim a, In-Young Choi a,b,c, Mary L. Michaelis d, Phil Lee a,c,⁎,1

a Hoglund Brain Imaging Center, University of Kansas Medical Center, Kansas City, KS 66160, USAb Department of Neurology, University of Kansas Medical Center, Kansas City, KS 66160, USAc Department of Molecular and Integrative Physiology, University of Kansas Medical Center, Kansas City, KS 66160, USAd Department of Pharmacology and Toxicology, University of Kansas, Lawrence, KS 66045, USA

⁎ Corresponding author at: Hoglund Brain Imaging CeStop 1052, University of Kansas Medical Center, Kansas913 588 9071.

E-mail address: plee2@kumc.edu (P. Lee).1 Please note the corresponding author's legal name chan

1053-8119/$ – see front matter © 2011 Elsevier Inc. Aldoi:10.1016/j.neuroimage.2011.02.039

a b s t r a c t

a r t i c l e i n f o

Article history:Received 27 October 2010Revised 1 January 2011Accepted 7 February 2011Available online 19 February 2011

Keywords:Axonal transportAlzheimer's diseaseMEMRITriple transgenic mouse modelT1 mapping

Impaired axonal transport has been linked to the pathogenic processes of Alzheimer's disease (AD) in whichaxonal swelling and degeneration are prevalent. The development of non-invasive neuroimaging methods toquantitatively assess in vivo axonal transport deficits would be enormously valuable to visualize early, yetsubtle, changes in the AD brain, to monitor the disease progression and to quantify the effect of drugintervention. A triple transgenic mouse model of AD closely resembles human AD neuropathology. In thisstudy, we investigated age-dependent alterations of the axonal transport rate in the triple transgenic mouseolfactory system, using fast multi-sliced T1 mapping with manganese-enhanced MRI. The data show thatimpairment in axonal transport is a very early event in AD pathology in these mice, preceding both depositionof Aβ plaques and formation of Tau fibrils.

nter, 3901 Rainbow Blvd, MailCity, KS 66160, USA. Fax: +1

ge from Sang-Pil Lee to Phil Lee.

l rights reserved.

© 2011 Elsevier Inc. All rights reserved.

Introduction

Neuronal cell function and viability critically depend on effectiveand timely axonal transport (AT), i.e. the process in whichintracellular cargoes including neurotransmitters, proteins andorganelles are trafficked through axons to and from the nerveterminals. Axonal transport deficits have been implicated in neuro-degenerative diseases such as Alzheimer's disease (AD) (Morfini et al.,2009; Wilson, 2008). Two major pathological hallmarks of AD,accumulation of β-amyloid (Aβ) and fibrillization of Tau, appear tocause axonopathy. Tau is a microtubule-associated protein thatstabilizes and is essential for maintaining normal axonal transportprocesses (Gotz et al., 2006). The dysregulation of axonal transportleads to synaptic dysfunction and synaptic/neuronal loss in animalmodels of AD. The presence of axonopathy in human AD has beensupported by findings of axonal swelling, accumulated mitochondrialcomponents in lysosomes, accumulation of vesicles in cell bodies, anddysfunctional synaptic vesicles at nerve terminals (Gotz et al., 2006;Stokin et al., 2005).

Increased axonopathy has been reported in animal models ofamyloidosis including the Tg2576 mice that over-expresse a mutanthuman amyloid precursor protein (APP) and APPmutantmice that alsoexpresse a human presenilinmutation (PS1) (Smith et al., 2007;Wirthset al., 2007). It has been suggested that the axonal transport deficits areoccurring in conjunction with the accumulation of insoluble Aβ andprior to Aβ plaque formation in the Tg2576 AD mouse model. Many ofthe transgenic AD mouse models exhibit axonal swelling that likelyinterferes with normal axonal transport of essential cellular compo-nents, mitochondria, and other vesicles. Mutations in the Tau gene alsoseriously disrupt microtubule function and lead to dystrophic axonalprocesses (Hutton et al., 1998; Spillantini et al., 1998). However, themechanisms by which both aging and various genetic insults disruptaxonal transport have not been clearly delineated (De Vos et al., 2008;Morfini et al., 2009; Schindowski et al., 2007). Therefore, understandingthe role of Aβ and Tau in the development of axonal transport deficits inAD brain is crucial for developing new strategies in early diagnosis andpossible treatments to slow progression of AD pathology.

Axonal transport rates have conventionally been measured in vitrousing radiotracers or fluorescent dyes in excised neurons. Theseconventional methods are invasive and cannot be used in vivo.Recently in vivo approaches to measuring axonal transport rateshave been reported using magnetic resonance imaging (MRI)methods. One commonly used form of MRI, manganese-enhancedmagnetic resonance imaging (MEMRI), involves application ofmanganese chloride (MnCl2) as a contrast agent to follow axonal

1287J. Kim et al. / NeuroImage 56 (2011) 1286–1292

transport. MEMRI measures the shortening of the longitudinalrelaxation rate constant (R1=1/T1) over time after administering anMnCl2 solution (Chuang and Koretsky, 2006; Kim et al., 2009).Manganese ion (Mn2+) is a paramagnetic, calcium analog that hasbeen used as a contrast agent inMRI. It is a trans-synaptic tracer that istaken up into neurons via voltage-gated calcium channels, packagedinto vesicles, and transported down the axon in a microtubule-dependent manner. After being released from pre-synaptic terminals,Mn2+ crosses the synapse, enters post-synaptic neurons, and isdistributed through interconnecting brain areas by selective antero-grade transport. In a series of papers, Pautler and co-workersestablished MEMRI as a viable neuronal tract tracing method thatallows for the measurement of axonal transport rates (Pautler et al.,1998; Pautler and Koretsky, 2002; Pautler, 2004). Given that aging isthe major risk factor for AD, it is worth noting that axonal transportrates in the CNS of rats, measured usingMEMRI, decrease with normalaging (Cross et al., 2008).

Currently, most MEMRI studies rely on T1-weighted (T1w) MEMRImethods to detect changes in signal enhancement from Mn2+

accumulation (Serrano et al., 2008; Smith et al., 2007). However, MRIcontrast in T1wMEMRImethods can only be optimal for a narrow rangeof Mn2+ concentration in the target tissue. In addition, quantification ofT1wMEMRI is not reliable for multi-session or longitudinal studies dueto variations in the signal intensity by other contributing factors such asflip angle and base line T1 values. Unlike T1wMEMRI, the parametric T1mapping method provides reliable quantification as the signal sensitiv-ity does not depend on the Mn2+ concentration. The T1 mappingmethod also allows the Mn2+ contrast comparison in multiple sessionsbecause this method does not require a normalization process. In thisstudy, we have used our previously-developed fast multi-slice Look–Locker sequence with multiple phase encodings per inversion pulse,which enabled T1 mapping in short acquisition time (Lee, 2007).

The olfactory system is particularly favorable for the study ofaxonal transport since Mn2+ can be delivered noninvasively. Thedelivery of Mn2+ into mouse olfactory epithelium can easily beachieved via injection of MnCl2 solution through the nasal pathway asit connects primary neurons in the epithelium to cortical projectionswithout necessitating transport of tracers across the blood-brainbarrier (Cross et al., 2004; Cross et al., 2008). Successful measurementof an axonal transport deficit has been achieved in the olfactorysystem of an APP transgenic mouse model of AD using the MEMRItechnique (Smith et al., 2007).

In this study, we have used an animal model of AD that expressesboth Aβ and Tau pathologies, a triple transgenic mouse model of AD(3×Tg-AD). The 3×Tg-AD is widely used as an animal model for humanAD because it expresses mutations in PS1M146V, APPSwe, and TauP301L,leading to bothAβ andTaupathologies (Oddoet al., 2003). The3×Tg-ADmice develop Aβ and Tau pathologic lesions in a sequential manner. Bymeasuring axonal transport deficits in this mouse model at differentages, we investigated the effects of the combination of Aβ and Tauaggregates on axonal transport at an early stage of the diseasedevelopment in these mice. We hypothesize that occurrence of earlyaxonal transport deficits is likely due to intraneuronal Aβ first, followedby Tau hyper-phosphorylation and fibrillization, leading to progressiveaxonal transport deficits. Using quantitative T1 mapping rather thanqualitative T1w imaging together with non-invasive MEMRI, wedemonstrate that these methods can be reliably employed to detectage-dependent axonal transport impairments.

Materials and methods

Animals

All the animals were handled in compliance with institutional andnational regulations and policies. The protocols were approved by theInstitutional Animal Care and Use Committee. 3×Tg-AD mice were

obtained from the mouse colony established at the University ofKansas from a breeding pair obtained from Dr. Laferla at theUniversity of California at Irvine. 3×Tg-AD mice harbor PS1M146V,APPSwe, and TauP301L and progressively develop both Aβ plaques andNFT pathology with accompanying neuronal death in brain regionssimilar to those seen in human AD. These animals developintracellular Aβ pathology in neocortex at 3 months of age, extracel-lular Aβ pathology in frontal cortex at 6 months, and Tau pathology at12 months. The presence of both Tau and APP transgenes in olfactorybulb (OB) has been confirmed byWestern blotting (Oddo et al., 2003).

Three age groups of 3×Tg-ADmice and age-matchedwild-type (wt)mice were studied: 2 months (2 mos), n=9 for wt and n=8 for 3×Tg-AD; 3 months (3 mos), n=8 for wt and n=4 for 3×Tg-AD; and15months (15 mos) of age, n=5 for wt and n=6 for 3×Tg-AD.

Manganese (Mn2+) administration

Animals were anesthetized with 4% isoflurane mixed in 4 L/min airand 1 L/min O2 for 5 min. Manganese chloride (MnCl2) solution wasadministered intranasally. A nasal lavage of 4 μL of isotonic MnCl2solution (160 mM) (Sigma-Aldrich, St. Louis, MO) was administered tothe left nostril using aHamilton syringe (Hamilton Company, Reno, NV).Animalswere then returned to a fresh cage on a heating pad, stimulatedusing amyl acetate for 15 min to enhance uptake of Mn2+ in theolfactory neurons, and allowed to recover fully before MRI sessions. Wechose unilateral administration of MnCl2 solution to reduce the dosageof the MnCl2 solution to limit possible toxicity and to facilitateconfirmation of the successful administration of MnCl2 by comparingsignal enhancements in the ipsilateral and contralateral turbinate.

Magnetic resonance imaging

All MR studies were performed using a 9.4 T Varian INOVA system(Varian Inc., CA), equippedwith a 12 cmgradient coil (40 G/cm, 250 μs).A 6 cm diameter Helmholtz volume transmit coil and a 7 mm diametersurface receive coil with de-tune capability were used for MR imaging.Animals were anesthetized and maintained with 1–1.5% isofluraneduring MRI sessions. Mice were under anesthesia for less than 40 minfor each MRI session. Core body temperature was maintained at 37 °Cusing a circulating hot water pad and a temperature controller (Cole-Palmer, NY). Respiration was monitored via a pressure pad under theanimal (SA Instruments,NY).MRdatawere acquiredbefore and1 h, 6 h,and24 hafter unilateral and intranasal administration ofMnCl2 solutionin four separate MRI sessions. These four MRI time points were chosenbasedon the timecourseof theR1 changes in the olfactory bulb (OB)andlateral olfactory tracts (LOT) after intranasalMnCl2 administration (datanot shown). The data showed that R1 in the OB changed linearly withtime from 1 h up to 6 h post MnCl2 administration. Therefore, 1 h and6 h time points were chosen to calculate axonal transport rates, toremove the effect of the initial transport delay from the turbinate to theOB, and to achieve high detectability of R1 changes in OB by allowingenough time for transported Mn2+ to accumulate in the OB. The 24 htime point was chosen to provide enough time for axonal transport ofMn2+ to the more distal LOT.

T1 maps were measured using a non echo planar imaging based,multi-slice T1 sequencemodified from the Look–Locker sequence (Lookand Locker, 1970) to acquire two phase encodings per inversion pulse(TR/TE=4/2 ms, FOV=2 cm, matrix=128×128, thickness=0.5 mm,flip angle=20°, 22 inversion times from 40–5470 ms, and acquisitiontime=8.5 min). B1 maps were measured to correct the effect of flipangle variations on T1 maps (TR/TE=200/3.7 ms, matrix=128×128,NEX=4, thickness=0.5 mm) (Pan et al., 1998). High resolution T1wspin-echo images were also acquired to visualize turbinate areas of theolfactory system (TR/TE=600/10 ms, NEX=2, matrix=256×256,thickness=0.45 mm, scan time=5min).

1288 J. Kim et al. / NeuroImage 56 (2011) 1286–1292

MEMRI data analysis

The T1 and B1 values of each pixel were calculated using a programwritten in IDL (RSI, CO). The R1 values were calculated based on thefollowing equation:

R1 =1T1

=1

T⁎1+

ln cos θð Þτ

ð1Þ

where T1⁎ is the apparent T1 obtained from the fitting of inversionrecovery signal intensities using the Levenberg–Marquardt algorithm,τ is the duration between successive excitations of the same slice, andθ is the flip angle. Ideally, if the exact flip angle at each pixel is known,the T1 can be computed from Eq. (1) after obtaining T1⁎ withoutfurther information. However, the B1 field inhomogeneity from theimperfect excitation pulse and the RF coil profile causes the flip angleto differ from the intended value. Therefore, B1 field maps were usedto obtain the actual flip angle value, θ in Eq. (1) in a pixel-by-pixelbasis. Region of interest (ROI) analysis was performed usingSTIMULATE software (Strupp, 1996). Axonal transport rate index(ATRI) of olfactory sensory neurons was calculated from the R1

changes in an ROI of olfactory bulb between 1 h and 6 h after MnCl2administration ((R1(6 h)−R1(1 h)) /5). ATRI at the LOT was estimatedfrom the time course of R1 changes in an ROI of the LOT between preand post 24 h MnCl2 administration ((R1(24 h)−R1(pre)) /24).

Mn2+ uptake in the turbinate was compared between wt and3×Tg-AD mice at 1 h post MnCl2 administration. The MR signalintensity in the spin-echo T1w images was measured from bothenhanced (left, L) and non-enhanced (right, R) sides of the turbinate.Unlike gradient-echo based T1 mapping methods, spin-echo T1wMEMRI does not suffer from the macroscopic susceptibility inducedsignal loss where the susceptibility effect is pronounced due to the airand tissue interface, such as the turbinate regions. The MR signalintensity ratio between two sides at 1 h post MnCl2 administrationprovided an index for Mn2+ uptake.

Trans-synaptic transmission efficiency index of Mn2+ was esti-mated from the ATRI at the OB and LOT. Since the concentration ofMn2+ at LOT is proportional to the concentration in the OB and theefficiency of synaptic transmission from the olfactory sensory neuronsto mitral cells in the OB glomeruli, the trans-synaptic transmissionefficiency indexwas calculated by the ratio of the ATRIs at LOT and OB.

Fig. 1.MEMRI of a 3×Tg-ADmouse. (A) T1w high resolution coronal MEMRI (top) and sagitta(i) turbinate; (ii) olfactory bulb (OB) and (iii) lateral olfactory tracks (LOT) at 6 h post MnClimages of (i) the turbinate at 1 h, (ii) the OB at 6 h, and (iii) the LOT at 24 h post MnCl2 adminillustrating the layered structure of the different cell types: the glomerular layer (GL); mitpanels A and B indicates the animal's right and “L” indicates the animal's left.Panel C image was adapted from Wikipedia, public domain.

Two-tailed Student t-testswere performed to compare groupmeansof ATRIs usingMicrosoft Excel. All data are presented asmean±SD. A p-value of less than 0.05 was considered statistically significant.

Results

High resolution MEMRI of the 3×Tg-AD mouse brain

The coronal Mn2+ enhanced T1w MR image (T1w MEMRI) clearlyshows structures and layers of the nasal and olfactory system of the3×Tg-AD mouse (Fig. 1A top). The T1w MEMRI was acquired at 6 hpost MnCl2 administration in a 5 min data acquisition time anddemonstrates the Mn2+ accumulation from the turbinate to the OB.Three major areas (i: turbinate; ii: OB; iii: LOT in Fig. 1A (top: coronalMRI, bottom: sagittal MRI)) were chosen to calculate axonal transportrates from the turbinate to the OB, and from the turbinate to the LOT.The first slice position (i in Fig. 1A) was placed where olfactorysensory neurons are located in the olfactory epithelium of theturbinate. The second slice position (ii) was placed where olfactoryneurons project to glomeruli in the olfactory bulbs. The third sliceposition (iii) was placedwhere olfactory output neurons (mitral cells)are receiving inputs from glomeruli and projecting to various targetsin the brain via olfactory tracts. Mn2+ accumulates following theolfactory neural circuits from olfactory sensory neurons to olfactorytracts, and MEMRI was measured at 1 h, 6 h, and 24 h post MnCl2administration through the left nasal cavity.

Mn2+ uptake in the turbinate at 1 h post MnCl2 administration

High resolution T1w MEMRI of the turbinate (Fig. 1B(i)) wasacquired at 1 h post MnCl2 administration. Unilateral MR signalenhancement in the left turbinate (marked by arrowhead) shows thestructure in detail, including nasal cavities. The signal enhancement inthe turbinate indicates Mn2+ uptake into the olfactory sensory neuronsduring the initial hour. There was no difference in signal enhancementbetween wt and 3×Tg-AD mice in any age, 2 mos, 15 mos or 18 mos ofage (p=0.57, n=5 for wt and n=6 for 3×Tg-AD).

Mn2+ transport in the OB at 6 h post MnCl2 administration

At 6 h post MnCl2 administration, the layered structure of the OBwas clearly identifiable with MR signal enhancement through Mn2+

l MRI (bottom) of a 3×Tg-ADmouse showing slice positions of MEMRI data acquisition:2 administration. (B) Corresponding T1w high resolution (78×78×450 μm3) transverseistration. (C) Extended view of the OB and a corresponding histological section of the OBral cell layer (ML); external plexiform layer (EPL); and granule cell layer (GrO). “R” in

1289J. Kim et al. / NeuroImage 56 (2011) 1286–1292

accumulation in the glomerular layer (GL) and the mitral cell layer(ML) in the high resolution T1w MEMRI (Fig. 1B(ii)) with theexpanded view of the figure in the left panel of Fig. 1C. The externalplexiform (EPL) and granule cell (GrO) layers did not show clearsignal enhancement. Cellular layers of the OB in mice are detailed in ahistological section figure (Fig. 1C, www.wikipedia.org) to illustratethe matching structure in the T1w MEMRI.

Quantitative measurement of axonal transport rate using T1-mapping

Fig. 2 shows representative T1maps (upper row, pseudo color) andT1w images (lower row, gray color) of the OB of 3 mos old wt (Fig. 2A)and 3×Tg-AD (Fig. 2B) mice at baseline, and 1 h, 6 h, and 24 h postMnCl2 administration through the nasal cavity. At 1 h post MnCl2administration, a slight T1 reduction was observed in the medial andlateral regions of the left OB in the T1 map of wt. The correspondingT1w signal enhancementwas also observed in the T1w image of wt. Nodiscernable changes were detectable in either the T1 maps or T1wMRIof 3×Tg-ADmice. At 6 h and 24 h, both wt and 3×Tg-ADmice showedclear T1 reductions (T1w MEMRI signal enhancement) at the medial,lateral and dorsal parts of the left OB. The time course of R1 wascalculated from the ROI placed mostly in the glomerular layer of thelateral part of the OB (square in Fig. 2A, top row). Group comparisonsof R1 in the OB of 3 mos old mice between wt and 3×Tg-AD mice areshown in Fig. 2C. R1 values of 3×Tg-AD mice were significantly lowerat 1 h (wt: 0.82±0.12 s−1, 3×Tg-AD: 0.69±0.04 s−1; p=0.043), 6 h(wt: 1.84±0.27 s−1, 3×Tg-AD: 1.31±0.0.28 s−1; p=0.005) and 24 h(wt: 1.76±0.36 s−1, 3×Tg-AD: 1.25±0.24 s−1; p=0.018) postMnCl2 administration, indicating lower axonal transport of Mn2+ in3×Tg-AD than that in wt mice.

Fig. 2. Quantification of the axonal transport rate. Two sets (A and B) of T1 maps (upper rowsAD (B)mice at baseline, and 1 h, 6 h, and 24 h post MnCl2 administration. The small square in(1 h)) /5) of the OB was calculated. The color bars indicate the range of T1 values from 0 s to 2animal's left; this orientation holds for all figures displayed. (C) Longitudinal relaxation radministration, quantified from the ROI shown in panel A. (*: p=0.04 for 1 h; **: p=0.005

Lower axonal transport rate in the OB of 3×Tg-AD

An average ATRI of the OB was calculated from the differencebetweenR1 values at 1 h and 6 h in the samemanner as in Fig. 2C. Groupcomparisonsbetween3 mos old 3×Tg-ADandwt showed that3×Tg-ADmice had a significantly lower axonal transport rate (p=0.034) (Fig. 3).Fig. 3A shows T1 maps and the corresponding T1w MEMRI of the OB(slice position ii in Fig. 1A)with the clear unilateral signal enhancementin the left OB of 3×Tg-AD at 6 h post MnCl2 administration. The axonaltransport rate remained lower in 3×Tg-ADmice compared towtmice at15 mos, although it did not reach statistical significance (p=0.087). Theoverall ATRI of 3×Tg-ADmice was lower by 27% at 3 mos and lower by54% at 15 mos compared to wt mice.

The effect of aging on the axonal transport ratewas assessed in threeage groups (2 mos, 3 mos, and 15 mos). The average ATRI fell with agefor bothwt and3×Tg-ADgroups although3×Tg-ADmice transport ratesfell more in all age groups. The difference between young mice (2 mosand 3 mos) and old mice was statistically significant in both groups(3×Tg-AD: p=0.001 for 2 mos vs. 15 mos, and p=0.028 for 3 mos vs.15 mos; wt: p=0.033 for 2 mos vs. 15 mos, and p=0.017 for 3 mos vs.15 mos). Age-dependent decreases of the ATRI from 2 mos of age were22% and 69% at 3 mos and 15 mos of age for 3×Tg-ADmice, respectively,while the decreases were 9% and 43% for wt mice.

Measurement of trans-synaptic axonal transport rate in the lateralolfactory tract

Fig. 4A shows T1 maps and the corresponding T1w MEMRI of theLOT (slice position iii in Fig. 1A) with the unilateral T1 reduction in theleft LOT of 3×Tg-AD at 24 h post MnCl2 administration. At earlier time

) with corresponding T1w images (lower rows) of the OB of 3 mos old wt (A) and 3×Tg-the T1map (top row rightmost) indicates the ROI inside of which the ATRI ((R1(6 h)−R1

.3 s. “R” in panel A (top row left most) indicates the animal's right and “L” indicates theate (R1) at the OB of 3 mos old mice at baseline, and 1 h, 6 h, and 24 h post MnCl2for 6 h; ***: p=0.02 for 24 h; n=8 for wt and n=5 for 3×Tg-AD).

Fig. 3. Axonal transport deficit in theOB. (A) T1map(top)with correspondingT1wMEMRI (bottom)at theOBof a3×Tg-ADmouse at2 mosof age.MEMRIwas performedat6 hpostMnCl2administration. A small square in the upper panel shows anROIwhere the ATRI in theOB ((R1(6 h)−R1(1 h))/5)was obtained. The color scale bar indicates the range of T1 values of 0–2.3 s.(B) Comparison of the ATRI in the OB between wt and 3×Tg-AD at 2 mos (n=9 for wt, n=8 for 3×Tg-AD), 3 mos (n=8 for wt, n=4 for 3×Tg-AD) and 15 mos (n=5 for wt, n=6 for3×Tg-AD) of age at the OB. (p=0.3 for 2 mos; *: p=0.034 for 3 mos; p=0.087 for 15 mos of age).

1290 J. Kim et al. / NeuroImage 56 (2011) 1286–1292

points (1 h or 6 h post MnCl2 administration), no signal enhancementwas observed in the LOT.

The ATRI was measured at the LOT in 2 mos, 3 mos and 15 mos old3×Tg-AD mice. Overall the axonal transport deficits in the LOT weresimilar to the transport deficits observed in the OB of 3 mos and15 mos mice. The results of the comparison indicate that a dramaticaxonal transport deficit occurred at 3 mos (p=0.039) of age andpersisted in older age mice (15 mos) (p=0.028) compared to age-matched wt mice (Fig. 4B). The difference between young mice(2 mos and 3 mos) and old mice was statistically significant in 3×Tg-AD mice (p=0.003 for 2 mos vs. 15 mos, and p=0.016 for 3 mos vs.15 mos), but there were no differences in wt mice (pN0.1). The age-dependent decreases of the ATRI from 2 mos 3×Tg AD mice were 14%and 54% at 3 mos and 15 mos of age, respectively. Trans-synaptictransmission efficiency index of Mn2+ in 3×Tg-AD mice did not differfrom that of wt mice in all age groups (pN0.19).

Fig. 4. Axonal transport deficits in the LOT. (A) T1 map (top) with corresponding T1wMEMRIpost MnCl2 administration. The circles indicate the LOT area with unilateral enhancement. Aof the LOT was obtained. The color scale bar indicates the range of T1 values of 0–2.3 s. (B) Cn=8 for 3×Tg-AD), 3 mos (n=8 for wt and n=5 for 3×Tg-AD) and 15 mos (n=5 for wt15 mos of age).

Discussion

Early axonal transport deficit in the 3×Tg-AD mouse brain

In this study, axonal transport deficits in 3×Tg-AD mice precededdeposition of Aβ plaques and neurofibrillary tangles. Our findings ofthe early axonal transport deficit in 3×Tg-AD mice at 3 mos of age arecoincident with the timing of intraneuronal Aβ accumulation, as3×Tg-AD mice are known to develop this pathological lesion at3–4 mos in the neocortex, extracellular Aβ pathology such as Aβplaque deposition at 5–6 mos in the frontal cortex, and the Taupathology at 12 mos of age (Oddo et al., 2003). In light of the recentdemonstration by Pigino and colleagues that intraneuronal Aβ,especially oligomers, can induce fast axonal transport deficits in thegiant squid axoplasm (Pigino et al., 2009), the intraneuronal Aβaccumulation in 3×Tg-AD mice at 3–4 mos of age may be a

(bottom) at the LOT of a 3×Tg-ADmouse at 2 mos of age. MEMRI was performed at 24 hsmall square in the upper panel shows an ROI in which the ATRI ((R1(24 h)−R1(Pre)) /24)omparison of the ATRI in the LOT between wt and 3×Tg-AD at 2 mos (n=9 for wt andand n=6 for 3×Tg-AD). (p=0.88 for 2 mos; *: p=0.039 for 3 mos; **: p=0.028 for

1291J. Kim et al. / NeuroImage 56 (2011) 1286–1292

contributor to the observed axonal transport deficits. Furthermore,the axonal transport deficits could be linked to cognitive impairmentobserved in 3×Tg-AD at the same age (Billings et al., 2005). Earlyaxonal transport deficits in 3×Tg-AD were also consistent withprevious findings in Tg2576 mice in which axonal transport deficitsoccurred at 6–7 mos of age, prior to Aβ plaque deposition (Smith et al.,2007).

The axonal transport deficits in 3×Tg-AD persisted to the later ageof 15 mos, the age range in which both extracellular Aβ and Taupathologies are known to be present (Oddo et al., 2003). Althoughextracellular Aβ and/or Tau pathology may not contribute to theinitial axonal transport deficits, it is highly probable that bothpathologies contribute to the further reduction of axonal transportrates at a later age compared to those at 3 mos of age.

Trans-synaptic axonal transport of Mn2+

The axonal transport in the LOT of 3×Tg-AD mice was alsoimpaired as observed in the OB of 3 mos and 15 mos. Note that theATRI in the LOT is an order of magnitude lower than that in the OB. TheATRI reflects the rate of Mn2+ accumulation in each region of interestand the amount of Mn2+ accumulation at a given time in each regiondepends on the distance from the Mn2+ uptake site (i.e., turbinate)and the availability of Mn2+ in the neurons for transport after theuptake. The longer axonal distances from the turbinate to the LOT andthe Mn2+ dilution factor at the OB as Mn2+ are transported from theOB to various brain regions have contributed to the lower accumu-lation of Mn2+ in the LOT.

The transport deficits at the LOT could be caused either by slowmovement of Mn2+ along the axons or by low efficiency of trans-synaptic transport of vesicles, since the ATRI at the olfactory tractsincludes contributions from (1) axonal transport into the OB, (2)synaptic transmission from the olfactory sensory neurons to mitralcells at glomeruli in the OB and (3) axonal transport to the LOT.Considering that there was no difference in trans-synaptic transmis-sion efficiency indices of Mn2+ (synaptic transmission from theolfactory sensory neurons to mitral cells) between 3×Tg-AD and wtmice in all age groups, the lower axonal transport rate index at theLOT in 3×Tg-AD is most likely due to slow axonal transport within theneurons.

Quantitative measurement of T1 mapping at high magnetic field, 9.4 T

We have achieved quantitative axonal transport rate measurementsin the mouse brain in vivo using our T1 mapping method. One of themajor challenges of T1mapping inmice is the longacquisition timewhenusing multiple TR/flip angle sequences or conventional single-sliceLook–Locker sequences. Another challenge is the possible imagedistortion and/or signal dropout when echo planar imaging sequencesare used to reduce the acquisition time. To overcome these problems,weused our fast multi-slice Look–Lockermethod that provided T1mappingin less than 9 min. By using B1 corrected quantitative T1 mapping, withan excellent image resolution (156×156×500 μm3) across the mousebrain we visualized the T1 shortening in the OB and the LOT.

Significantly lower dose of MnCl2 for MEMRI

Mn2+ has proven to be useful in monitoring Ca2+ influx andmeasuring axonal transport rates in biological research (Serrano et al.,2008; Smith et al., 2007); however it is important to note thatmanganese is neurotoxic at high concentrations. Since the toxicity ofmanganese is associated with high concentrations, it is imperative toreduce the dose to avoid possible neurotoxicity, especially in longitu-dinal studies. In this study,we could lower the concentration ofMnCl2 to20 times lower (b5% of the dosages) than those used in previous studies(Serranoet al., 2008; Smith et al., 2007)with a sufficientlyhigh contrast-

to-noise ratio (CNR). By allowing 6 h ofMn2+ accumulation in the OB inconjunction with our T1 mapping method and high magnetic field, thedetection sensitivity of MRI signal changes was greatly enhanced. Thus,wewere able to obtain robust detection ofMEMRI signal changes at thislow total dose of Mn2+ and to quantify the axonal transport ratesreliably.When animals were subjected to multipleMEMRI sessions, theaxonal transport rates were consistent, and we did not observe anybehavioral changes, suggesting no observable ill effects from Mn2+ atthe current dose.

Potential limitation of axonal transport measurement using MEMRI andfuture directions

Since MEMRI measurements rely on Mn2+ accumulation and itsconcentration in the OB, the axonal transport rate measured usingMEMRImethodsmay include other contributing factors, such as uptakeof Mn2+ in the olfactory sensory neurons in the turbinate, packaging ofMn2+ into the vesicles, and cargo binding to themotor protein, kinesin.In this study, we did not observe any difference in signal enhancementin the turbinate area between 3×Tg-AD and wt mice in all age groupsdemonstrating that uptake of Mn2+ in the olfactory sensory neurons islargely identical. However, MEMRI does not provide any informationregarding packaging of Mn2+ and cargo binding. Therefore, the axonaltransport indices measured using MEMRI should be interpreted as anindex for overall axonal transport integrity rather than a simple rate ofaxonal cargo movement. Considering the fact that there are multipletargets for disturbing axonal transport in neurodegenerative diseasesincluding AD, axonal transport rate measurement using MEMRI in vivocan be used to assess overall integrity of axonal transport system thatmay be relevant to cognitive and behavioral outcomes.

The current findings of the axonal transport deficit could becorrelated with biochemical and histochemical analysis results toassociate with the presence of Aβ and Tau pathologies. Further study isneeded to understand the mechanisms responsible for early axonaltransport deficits in 3×Tg-AD mice. In addition, the longitudinal axonaltransport measurements should be possible by taking advantage oflower MnCl2 dosage and T1 mapping techniques to investigate theaxonal transport deficits in disease progression and the effect of thedrugtreatment in the same cohorts of animals.

Conclusion

This study demonstrates that high resolution quantitative MEMRIis a valuable method for non-invasive measurements of axonaltransport in the mouse brain. The combination of a novel quantitativeT1 mapping MEMRI technique, a high sensitivity RF coil set and highmagnetic field of 9.4 T, provided a robust measurement of R1 changesin OB and other parts of the brain as well as a visualization of cellularlayers of the OB structure with high signal-to-noise ratio. Thus, this invivoMEMRI approach should enable us to investigate axonal transportdeficits in various transgenic mouse models of AD to evaluate thedisease progression and the effect of drug treatments. This study alsoreports quantitative measurements of very early, yet subtle, axonaltransport deficits in olfactory systems of 3×Tg-AD mice. Our datasuggest that accumulation of intra-neuronal Aβ rather than the extra-cellular Aβ or abnormal Tau may initially lead to the axonal transportimpairment. However, further studies are needed to directly linkintra-neuronal Aβ and axonal transport deficits and to assess theeffect of extracellular Aβ and Tau pathologies on axonal transport.

Acknowledgments

This study was supported in part by the Alzheimer's Association(NIRG-07-60405 to Dr. Lee) and the National Institutes of Health (R21AG027419 to Dr. Michaelis and P30 HD002528 to the University of

1292 J. Kim et al. / NeuroImage 56 (2011) 1286–1292

Kansas). The Hoglund Brain Imaging Center is supported by C76HF00201 and a generous gift from Forrest and Sally Hoglund.

References

Billings, L.M., Oddo, S., Green, K.N., McGaugh, J.L., LaFerla, F.M., 2005. IntraneuronalAbeta causes the onset of early Alzheimer's disease-related cognitive deficits intransgenic mice. Neuron 45, 675–688.

Chuang, K., Koretsky, A.P., 2006. Improved neuronal tract tracing using manganeseenhanced magnetic resonance imaging with Fast T1 mapping. Magn. Reson. Med.55, 604–611.

Cross, D.J., Minoshima, S., Anzai, Y., Flexman, J.A., Keogh, B.P., Kim, Y., Maravilla, K.R.,2004. Statistical mapping of functional olfactory connections of the rat brain invivo. Neuroimage 23, 1326–1335.

Cross, D.J., Flexman, J.A., Anzai, Y., Maravilla, K.R., Minoshima, S., 2008. Age-relateddecrease in axonal transport measured by MR imaging in vivo. Neuroimage 39,915–926.

De Vos, K., Grierson, A.J., Ackerley, S., Miller, C.C.J., 2008. Role of axonal transport inneurodegenerative diseases. Annu. Rev. Neurosci. 31, 151–173.

Gotz, J., Ittner, L.M., Kins, S., 2006. Do axonal defects in tau and amyloid precursorprotein transgenic animals model axonopathy in Alzheimer's disease? J. Neuro-chem. 98, 993–1006.

Hutton, M., Lendon, C.L., Rizzu, P., Baker, M., Froelich, S., Houlden, H., Pickering-Brown, S.,Chakraverty, S., Isaacs, A., Grover, A., Hackett, J., Adamson, J., Lincoln, S., Dickson, D.,Davies, P., Petersen, R.C., Stevens,M., deGraaff, E.,Wauters, E., van Baren, J., Hillebrand,M., Joosse, M., Kwon, J.M., Nowotny, P., Che, L.K., Norton, J., Morris, J.C., Reed, L.A.,Trojanowski, J., Basun, H., Lannfelt, L., Neystat, M., Fahn, S., Dark, F., Tannenberg, T.,Dodd, P.R., Hayward,N., Kwok, J.B., Schofield, P.R., Andreadis, A., Snowden, J., Craufurd,D., Neary, D., Owen, F., Oostra, B.A., Hardy, J., Goate, A., van Swieten, J., Mann, D., Lynch,T., Heutink, P., 1998. Association of missense and 5′-splice-site mutations in tau withthe inherited dementia FTDP-17. Nature 393, 702–705.

Kim, J., Choi, I.-Y., Michaelis, M.L., Lee, S.-P., 2009. Early impaired axonal transport in atriple transgenic mouse model of Alzheimer's disease. Proc Intl Soc Mag ResonMed18, 540.

Lee, S.-P., 2007. Quantitative assessment of axonal transport integrity in the mousebrain in vivo using MEMRI at 9.4 T. Proc Intl Soc Mag Reson Med 15, 2441.

Look, D.C., Locker, D.R., 1970. Time saving in measurement of NMR and EPR relaxationtimes. Rev. Sct. Instrum. 41, 250–251.

Morfini, G.A., Burns, M., Binder, L.I., Kanaan, N.M., LaPointe, N., Bosco, D.A., Brown Jr., R.H.,Brown, H., Tiwari, A., Hayward, L., Edgar, J., Nave, K., Garberrn, J., Atagi, Y., Song, Y.,Pigino, G., Brady, S.T., 2009. Axonal transport defects in neurodegenerative disease.J. Neurosci. 29, 12776–12786.

Oddo, S., Caccamo, A., Shepherd, J.D., Murphy, M.P., Golde, E., Kayed, R., Metherate, R.,Mattson, M.P., Akbari, Y., LaFerla, F.M., 2003. Triple-transgenic model of Alzheimer'sdisease with plaques and tangles: intracellular Ab and synaptic dysfunction. Neuron39, 409–421.

Pan, J.W., Twieg, D.B., Hetherington, H.P., 1998. Quantitative spectroscopic imaging ofthe human brain. Magn. Reson. Med. 40, 363–369.

Pautler, R.G., Silva, A.C., Koretsky, A.P., 1998. In vivo neuronal tract tracing usingmanganese-enhanced magnetic resonance imaging. Magn. Reson. Med. 40, 740–748.

Pautler, R.G., Koretsky, A.P., 2002. Tracing odor-induced activation in the olfactory bulbsof mice using manganese-enhanced magnetic resonance imaging. NeuroImaging16, 441–448.

Pautler, R.G., 2004. In vivo, trans-synaptic tract-tracing utilizing manganese-enhancedmagnetic resonance imaging (MEMRI). NMR Biomed. 17, 595–601.

Pigino, G., Morfini, G., Atagi, Y., Deshpande, A., Yu, C., Jungbauer, L., LaDu, M., Busciglio,J., Brady, S., 2009. Disruption of fast axonal transport is a pathogenic mechanism forintraneuronal amyloid beta. Proc. Natl Acad. Sci. USA 106, 5907–5912.

Schindowski, K., Belarbi, K., Buee, L., 2007. Neurotrophic factors in Alzheimer's disease:role of axonal transport. Genes Brain Behav. 7, 43–56.

Serrano, F., Deshazer, M., Smith, K.D.B., Ananta, J.S., Wilson, L.J., Pautler, R.G., 2008.Assessing transneuronal dysfunction utilizing manganese-enhancedMRI (MEMRI).Magn. Reson. Med. 60, 169–175.

Smith, K.D.B., Kallhoff, V., Zheng, H., Pautler, R.G., 2007. In vivo axonal transport ratesdecrease in a mouse model of Alzheimer's disease. Neuroimage 35, 1401–1408.

Spillantini, M.G., Murrell, J.R., Goedert, M., Farlow, M.R., Klug, A., Ghetti, B., 1998.Mutation in the tau gene in familial multiple system tauopathy with preseniledementia. Proc. Natl Acad. Sci. USA 95, 7737–7741.

Stokin, G.B., Lillo, C., Falzone, T.L., Brusch, R.G., Rockenstein, E., Mount, S.L., Raman, R.,Davies, P., Masliah, E., Williams, D.S., Goldstein, L.S., 2005. Axonopathy and transportdeficits early in the pathogenesis of Alzheimer's disease. Science 307, 1282–1288.

Strupp, J.P., 1996. Stimulate: a GUI based fMRI analysis software package. Neuroimage3, S607.

Wilson, R.J., 2008. Towards a cure for dementia: the role of axonal transport inAlzheimer's disease. Sci. Prog. 91, 65–80.

Wirths, O., Weis, J., Kayed, R., Saido, T.C., Bayer, T.A., 2007. Age-dependent axonaldegeneration in an Alzheimer mouse model. Neurobiol. Aging 28, 1689–1699.