Age-related decrease in axonal transport measured
by MR imaging in vivo
Donna J. Cross,a,dJennifer A. Flexman,cYoshimi Anzai,b
Kenneth R. Maravilla,a,band Satoshi Minoshimaa,b,c,d,⁎
aWashington National Regional Primate Center, Washington, USA
bDepartment of Radiology, University of Washington, Seattle, Washington, USA
cDepartment of Bioengineering, University of Washington, Seattle, Washington, USA
dNeuroscience Program, University of Michigan, Ann Arbor, MI, USA
Received 4 April 2007; revised 18 July 2007; accepted 17 August 2007
Available online 31 August 2007
Axonal transport is a crucial process for neuronal homeostasis and cell
functions. In vitro studies have indicated transport rates decrease with
age. Disruption of axonal transport has been implicated in age-
associated neurodegenerative disorders. We hypothesized that aged
rats would show decreased transport in the brain, which could be
measured using in vivo manganese-enhanced MR imaging (Mn-MRI)
and parametric estimation. Serial T1-weighted images were obtained
at pre- and post-administration of MnCl2 in rats scanned long-
itudinally (n=4) and in a separate aged group (n=3). Subtraction
analysis was performed for group-wise statistical comparison on a
pixel-by-pixel basis. Change in intensity over time was plotted for the
olfactory bulb and anterior and posterior olfactory tract. Bulk
transport of material was estimated over an initial 72 h. Tracer kinetic
estimation of time–intensity data, based on a mass transport model,
used intensity change in the bulb as input function for subsequent
changes in the tract. Time to the peak of Mn2+flow was estimated for
both anterior and posterior tracts. Results indicated age-related
decreases in axonal transport rate and bulk transport of material in
the olfactory tract of living rat brains. Longitudinally scanned, mid-
age group was decreased by 58% and the aged group by 71% of young
rate (neuronal transport=4.07±1.24 mm/h, 1.72±0.89 mm/h, and
1.16±0.18 mm/h for young, mid-age, and aged, respectively).
Neuronal transport rate decreases correlated with increased age. The
use of kinetic analysis combined with dynamic manganese enhanced
MR imaging provides a unique opportunity to study this important
© 2007 Elsevier Inc. All rights reserved.
Axonal transport is crucial to normal neuronal cell function and
viability. Protein “cargos” synthesized in the cell body are
transported in an anterograde manner via molecular ‘motors’ to
the synapse for incorporation into membranes, cytoskeletal
functions, or released into the synaptic cleft (Ochs, 1972; Schwartz,
1979). Retrograde transport of substances taken up through
endocytosis, such as trophic factors will in turn direct further
protein synthesis in the cell body. Although axonal transport has
been studied using in vitro cell cultures and in vivo in peripheral
neuronal models such as sciatic nerve preparations, studies of
cortical axonal transport are limited, and differences between
peripheral and CNS transport rates has been suggested (Fibiger
et al., 1972; Jacob and O’Donoghue, 1995). Axonal transport is a
dynamic neuronal process that may be influenced by many factors
such as energy availability, neuronal cell type and type of cargo
being transported (Ochs, 1972; Verdu et al., 2000). Therefore,
extrapolation of results from in vitro studies or peripheral nerve
preparations to living brains may not be entirely accurate. Previous
studies mainly performed in peripheral neurons indicate that axonal
transport rates decrease with age under normal conditions (Ochs,
1973; Stromska and Ochs, 1982; Brunetti et al., 1987; McQuarrie
et al., 1989; Frolkis et al., 1997; Verdu et al., 2000; Uchida et al.,
2004). There have also been indications that several age-associated
neurodegenerative disorders, such as Alzheimer’s disease may have
disruption of axonal transport at a very early pathophysiological
phase of the disease process (Cash et al., 2003; Pigino et al., 2003;
Stokin and Goldstein, 2006a,b). Thus, the investigation of normal
age-related changes in transport is critical to the understanding of
pathophysiology of age-related disorders.
shortening effect on the relaxation constant T1. The resultant image
shows increased signal intensity in cortical regions containing the
manganese. It is a transsynaptic tracer that is taken up into neurons
via voltage-gated Ca2+channels, packaged into vesicles, and
transported down the axon in a microtubule-dependent manner
NeuroImage 39 (2008) 915–926
⁎Corresponding author. Department of Radiology, University of
Washington, 1959 N.E. Pacific Street, BB201c, Box 357115, Seattle, WA
98195-7115, USA. Fax: +1 206 543 6317.
E-mail address: firstname.lastname@example.org (S. Minoshima).
Available online on ScienceDirect (www.sciencedirect.com).
1053-8119/$ - see front matter © 2007 Elsevier Inc. All rights reserved.
After being released from presynaptic terminals, Mn2+crosses the
synapse, enters post-synaptic neurons, and is distributed through
interconnecting brain areas by selectively anterograde transport
(Sloot and Gramsbergen, 1994; Pautler and Koretsky, 2002).
Previously, our lab has investigated functional transport of
manganese in the olfactory system of normal rats (Cross et al.,
2004) and alterations in function transport after injury (Cross et al.,
The rodent olfactory system has been studied extensively with
classic track tracing investigations using agents such as wheat-germ
agglutinin conjugated to horseradish peroxidase (Itaya, 1987) and
radioisotopes of metal ions (Tjalve and Henriksson, 1999; Brenne-
man et al., 2000). It isa particularly favorable system to study due to
its connections from primary neurons in the epithelium to cortical
projections without necessitating transport of tracers across the
blood–brain barrier. The injection of manganese onto olfactory
epithelium via a small catheter is completely non-invasive and the
amount required for enhanced signal on MRI is below levels shown
to be neurotoxic (Brenneman et al., 2000; Henriksson and Tjalve,
We hypothesized that normal aged rats would show decreased
axonal transport in the CNS, which could be measured using in vivo
manganese-enhanced serial MR imaging (Mn-MRI) and parametric
estimation of transport rates. In vivo imaging of axonal transport
allows assessment of axonal processes in a quantitative manner.
Once established, this technique could be applied potentially to
human studies. In addition, the olfactory system has been shown to
be particularly vulnerable to both age-related changes and disease
and environmentally originated pathologies (Enwere et al., 2004;
Albers et al., 2006; Attems and Jellinger, 2006; Luzzi et al., 2006;
Barrios et al., 2007). Successful development of this non-invasive
technique for humans could pioneer a new method for study of
olfactory dysfunction in aging and disease.
Materials and methods
MnCl2administration and MR scanning
Four male young (4 months) and three aged breeder (24–
25 months, Charles River Labs) Sprague-Dawley rats had free
access to food and water and were kept on a normal 12-h light/dark
cycle. All procedures were approved by the University of
Washington Animal Care Committee. The four young rats were
scanned on 2 separate occasions at 4 months old and 13 to
17 months old (9 to 13 months apart) for a longitudinal assessment
of decreased transport. In addition, to control for potential con-
founding effects from cumulative manganese neurotoxicity, a group
of aged rats was scanned in only one session when they were very
aged (24–25 months).
Rats were anesthetized with isoflurane gas during pre- and post-
administration of MnCl2for serial MR scans of 1, 3–6, 11–12, 24,
36, 48, 72 h. In addition, rats were scanned for a long-term time
point 7–10 days post administration. After the pre-administration
scan, before recovering from anesthesia, rats received 8–10 μl
administration of 1 M MnCl2to the right nasal cavity via Hamilton
syringe with a 24 gage I.V. catheter attached. This procedure was
described in detail in our previous study (Cross et al., 2004).
For this study, serial T1-weighted images were acquired on a
1.5 T MR scanner (Signa, GE, Milwaukee, WI) using a rat brain
volume coil and a 3D-SPGR (3-D spoiled gradient recalled
acquisition in a steady state: GRASS) pulse sequence. The original
256 by 160 acquisition matrix was reconstructed to 256×256 image
matrix using zero-filled interpolation pixels with a 0.5 mm slice
thickness in 88 continuous slices. Original image voxel size was
0.273×0.273×0.5. The following parameters were used; FOV=
7×4 cm; echo time (TE)=6.8 ms; repetition time (TR)=15 ms; flip
angle 45°; repetition 4 NEX. Total time under anesthesia for eachrat
was about 10 min per scan.
As described previously (Cross et al., 2004), image processing
used fully automated programs that are incorporated into a software
library for neuroscience image analysis (NEUROSTAT, University
of Washington). Serial image sets for each subject were coregistered
to a common orientation using rigid-body transformation. All co-
registered image sets were then registered to the Rat Atlas in
Stereotaxic Coordinates (Paxinos and Watson, 1998). In stereotactic
registration processing, the coregistered image sets for each subject
were averaged, and the averaged image set was matched to an MR
stereotactic template brain. The estimated 9 affine parameters were
applied to individual scans, permitting consistent stereotactic trans-
formation for scans from the same subject. All image sets were
els, 140 slices, voxel size of 0.2 mm). The third step of the process
intensity on scans obtained at different time points. Normalization
was comprised of within-subject normalization across all scans and
followed by linear normalization across all rats using a modified
stochastic algorithm to match the rat brain template intensity
(Minoshima et al., 1993). These image processing algorithms were
described previously in greater detail (Cross et al., 2004).
Z-statistic mapping of post administration transport
Subtraction analysis allows a group-wise paired statistical com-
manganese enhancement. Using this algorithm, pre-administration
scans were subtracted from post-administration scans across sub-
jects at each time point. One-sample t-statistic values were calcu-
lated across rats for each subtracted pixel value and converted to
z-statistic maps using a probability integral transformation (Friston
et al., 1989). The resultant z-statistic maps represented the extent
and significance of manganese transport within the brain averaged
across each group. These mapping techniques were used in the
To investigate manganese transport within the lateral olfactory
tract at different time points following intranasal administration, we
examined z-statistic maps created for each post-administration time
point. After all image sets were standardized into the stereotactic
space, and pixel intensities were normalized, group-wise z-statistic
subtraction maps were generated. The baseline pre-administration
image was subtracted statistically from image sets at each time point
better localize the peak uptake.
The threshold for significant enhancement was determined using
the theory of a smoothed stochastic process (Worsley et al., 1996).
To control a Type I error rate at p=0.05 adjusted for multiple pixel
comparison, the number of resolution elements (Worsley et al.,
1992) was estimated from the number of pixels and smoothness of
the z-statistical map (Friston et al., 1990), and a critical z-threshold
916D.J. Cross et al. / NeuroImage 39 (2008) 915–926
was estimated using Euler’s characteristics (Worsley et al., 1992).
The estimated critical threshold of Z=4.0 (controlling p=0.05
adjusted for multiple pixel comparisons) was used in this analysis.
Volume of interest (VOI) analysis
The time course of manganese uptake in the olfactory bulb and
subsequent transport down the lateral olfactory tract was measured
using conventional volume of interest (VOI) analysis. VOIs in the
bulb and in the anterior and posterior olfactory tract (+4.6 mm
from and at the bregma landmark slice level, respectively) were
determined using stereotactic coordinates (Paxinos and Watson,
1998) placed on transformed, pixel–intensity-normalized MR
image sets. A schematic of the main olfactory pathways in this
study has been adapted from a previous publication (Fig. 1) (Cross
et al., 2006b). This diagram illustrates the relative connectivity of
the VOIs and indicates that the VOIs are not separated by synaptic
connections. The olfactory bulb contains the cell bodies of the
neurons that have axons projecting through the lateral olfactory
tract. Anterior and posterior VOIs are within that tract.
After subtraction of pre-administration VOI value, significant
differences in intensity were determined by one-sample t-statistic
test for young compared to mid-age subjects and two-sample for
youngcompared to agedgroup(p≤0.05). Thechangein normalized
intensity over time was plotted for each VOI for each subject. In
addition, change in intensity at each time point was group-averaged
for each VOI. The plots were used in the subsequent curve-fitting
To estimate bulk transport of material from bulb to olfactory
tract, area under the curve for the initial 72 h was calculated for all
three VOIs. For each subject, the average VOI intensity from Time
point A to Time point B was multiplied by change in time from A
Fig. 1. Main rat olfactory routes of Mn2+delivery after intranasal administration. Diagram illustrates some of the relevant structures showing increased signal
intensity following MnCl2administration. Mn transport relevant to the current study is indicated by red arrows. Blue arrows indicate other transport routes.
Relative VOI locations are also indicated. This diagram was adapted from a previous publication (Cross et al., 2006b).
917 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
to B and summed over all time points to get the total change of
intensity over the first 72 h. Bulk manganese transport from bulb to
anterior and posterior olfactory tract VOIs was calculated as a ratio
of total intensity change in of tract VOI divided by bulb. In addition,
transport of manganese from anterior to posterior tract was
calculated as posterior/anterior ratio and expressed as a percent.
Tracer kinetic estimation of axonal transport speed
To support the results indicated by pixel-by-pixel statistical
mapping and bulk transport estimation, we applied tracer kinetic
estimation to the olfactory tract dynamic changes in intensity. In this
study, the Mn2+ion is taken up by the primary neurons of the
olfactory nerve and transported to the olfactory bulb. Axons of the
olfactory nerve neurons form synaptic connections in the glomeruli
of the bulb with the mitral and tufted cells. The axons of the mitral
cells comprise the lateral olfactory tract (Fig. 1). To distinguish
tracer uptake by neurons versus transport changes, we used tracer
intensity as an input function to the tract. The VOI were carefully
(the origin of the tract input) as well as 2 locations in the tract; one
more anterior and the other at the most posterior portion of the tract.
The changes in VOI signal intensity were subtracted from baseline
intensity and plotted against time over all the scan sessions. To
estimate axonal transport rates, a curve-fitting analysis of time–
intensity data was performed that was derived from the theory of
longitudinal dispersion and travel time of contaminants in rivers and
bulb was convolved with a mass transport function and used as the
lateral olfactory tract and time to the peak of bulk Mn2+flow was
estimated for both VOI regions (anterior and posterior) (Fig. 2). The
first-order lossterm andglobal scalingwere first estimated usingthe
group-averaged data, and then fixing these parameters, individual
values for dispersion coefficient and time-to-peak flow were
estimated for each region of each subject. The rate of axonal
transport of Mn2+along the olfactory tract axons was estimated by
template. This value was then scaled to the individual animal (to
adjust for differences in brain size between young and aged rats).
library files that were automatically generated when individual
brains were transformed to the atlas template.
Statistical mapping indicates Mn2+transport differences with age
Group-wise pixel-by-pixel statistical mapping indicated age-
related decreases in manganese transport to the olfactory tract both
anterior and posterior regions. Peak z-values for both anterior and
posterior VOIs were calculated across time points from 1 h to 48 h.
Using the threshold Z=4.0 as indicating significantly increased
enhancement, the young group showed significant manganese
transport through the posterior tract as early as 11–12 h (Z=6.0 and
Z=4.7, for anterior and posterior, respectively) (Fig. 3). When the
young rats were re-scanned at mid-age, significant changes of
Fig. 2. Curve fitting of raw normalized intensity for peak flow estimates in olfactory tract VOIs. Vertical axis indicates the normalized (unsubtracted) raw
intensity data from a single subject. Horizontal axis indicates time post administration of MnCl2. Bulb uptake was used as input to the olfactory tract VOI
parameter estimates. Actual data are shown as points and fitted estimates at each time point are shown as dashed lines (black and gray for anterior and posterior,
respectively). Peak flow parameter in a VOI was estimated with respect to bulb uptake. In this example from a mid-age rat, peak flow for the anterior tract was
estimated at the time of 981 min post-administration. Posterior tract peak flow was estimated at the time of 1148 min. post. Axonal transport estimate between the
2 VOIs was calculated by dividing the distance between them by the difference in time (6.14 mm/(1148−981) min×60 min/h=2.21 mm/h).
918 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
intensity in the olfactory tract was not seen at the early time point.
However by 24 h, significant enhancement was observed through
the posterior tract. Interestingly, in the aged group, significant
enhancement was indicated in the anterior tract by 24 h, but did not
reach statistical significance in the posterior tract VOI (Z=3.5).
Peak z-values in the olfactory tract from anterior to posterior were
measured at 24 h on a slice-by-slice basis (Fig. 4). Peak z-value
decreased with distance along the tract (Pearson’s correlation
coefficient=−0.93) and fell below the statistical threshold, Z=4.0,
at 5.8 mm from anterior VOI (bregma +0.4 mm).
Bulk material transport to olfactory tract is decreased in aged rats
Averaged change in MR intensity over time was plotted for
each group. While the magnitude of bulb enhancement was similar
across groups (Fig. 5A), enhancement over time in the anterior and
posterior tract was reduced in mid-age and aged groups (Figs. 5B
and C). To examine this difference quantitatively, bulk transport of
manganese from bulb to the anterior and posterior olfactory tract
was estimated as area under the curves over the first 72 h and then
the ratio of tract change in intensity was divided by bulb change in
intensity. Young rats showed significantly greater conservation of
intensity between the bulb and both anterior and posterior tract as
compared to aged group (anterior 71±8%, 56±8% and posterior
68±9%, 40±3% for young and aged, respectively, p≤0.05). In
addition, of the material that was transported into the anterior tract,
there was a significantly greater conservation to the posterior tract
in young compared to aged over the initial 72 h (97±7% to 75±
14%, p≤0.05) (Fig. 6). Mid-age rats showed middle range
conservation (57±5%, 49±15% and 82±19% for anterior/bulb,
posterior/bulb and posterior/anterior respectively).
Aged rat Mn2+uptake into bulb slower but not decreased
The change in signal intensity that was measured over the
scanning sessions in the bulb was significantly faster. Closer
Fig. 3. Mn2+through the lateral olfactory tract is delayed with aging. Setting a statistical threshold of Z=4.0 (after Bonferroni's correction for comparison across
multiple pixels), group-wise statistical maps of changes in MR intensity indicate significant peak enhancement in both anterior (blue arrow) and posterior (purple
arrow) at the EARLY (11–12 h) time point. Mid-age group statistical maps indicate sub-threshold enhancement of both anterior and posterior VOI until the 24-h
time point. In contrast, aged group peak z-value exceeded the threshold by 24 h in the anterior VOI, but was still sub-threshold in the posterior VOI. By 36 h post
MnCl2administration, aged group posterior VOI reached statistical threshold. Statistical maps are shown in coronal slices superimposed onto a template MRI for
localization purposes. Anterior tract=+4.6 mm, posterior tract=+0.2 mm from bregma landmark.
919 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
examination of the average change in intensity over the initial 72 h
showed that young subjects reached peak intensity in the bulb at
around 11–12 h and then decreased over time (Fig. 7A). In contrast,
both mid-age and aged subjects showed peak intensity in the bulb
that is closer to 24 h before starting to decrease. The only difference
in intensity magnitude that reached significance was between young
and aged groups at 11–12 h (p≤0.05).
Peak intensity versus peak flow estimation
The same trend was apparent in the anterior and posterior tract
where the young group reached peak intensity at 24 h and mid-age
and aged groups were closer to 36 h (Figs. 7B and C). Curve-fitting
parameter estimation (Fig. 2) was used to estimate the peak flow in
each VOI with the individual bulb change in intensity as the input.
Curve-fitting estimation indicated a significant increase in time
required to reach peak flow in both anterior and posterior tracts of
rats scanned at mid-age compared to when young (Fig. 8). Aged
rats showed even greater increase in time to peak flow compared to
young. These results were independent of bulb uptake differences.
Neuronal transport rate decreased in aged rat brains
Using individual estimates of time to peak flow to anterior and
posterior olfactory tract, rate of neuronal transport of Mn2+was
estimated. Both mid-age, and aged rats had significantly decreased
rates of Mn2+transport as compared to young rats. The long-
itudinally scanned, mid-age group was decreased by 58% and the
aged group was decreased by 71% of young rate (neuronal
transport=4.07±1.24 mm/h, 1.72±0.89 mm/h, and 1.16±
0.18 mm/h for young, mid-age, and aged rats, respectively).
Neuronal transport rate decreases correlated with increased age
(Pearson’s correlation coefficient=−0.82) (Fig. 9).
The results of this study indicated age-associated axonal
transport deficits including longitudinal differences in living rat
brains. Longitudinal decrease in brain axonal transport rates of
aged rats has not been reported previously. Investigations of axonal
transport in the brain are limited, in part due to the relative
inaccessibility of living brain neuronal tracts and the technical
challenges inherent in studying a dynamic, sub-cellular process.
Axonal transport rates have been shown to decrease with age under
normal conditions in peripheral neurons (Ochs, 1973; Stromska
and Ochs, 1982; Brunetti et al., 1987; McQuarrie et al., 1989;
Frolkis et al., 1997; Verdus et al., 2000; Uchida et al., 2004). Our
investigation of mid-age subjects rescanned after 9–13 months
showed a significant decrease of axonal transport in the lateral
olfactory tract. Very aged rats scanned for only one session showed
decreased axonal transport as well as decreased bulk transport of
material over the first 72 h. These results were shown initially by
group-wise z-statistical analysis as well as evaluation of changes in
MR intensity over time in the anterior and posterior olfactory tract.
To support these findings, kinetic analysis of time to peak flow in
tract VOI was used for more precise quantification of transport
rates that were independent of bulb uptake delays. The technique
used in this study has the great advantage of permitting evaluation
of this critical neuronal process in a living brain in a longitudinal
manner and non-invasively.
Fig. 4. Z-value decreases with distance from anterior olfactory tract at 24 h post-administration of MnCl2. Vertical axis indicates peak z-value in lateral olfactory
tract. Horizontal axis indicates distance (mm) from anterior tract VOI. Solid black line shows statistical threshold for significant enhancement on group-averaged
statistical maps. Black arrow indicates position of posterior VOI at +0.0 bregma. Distances are scaled to the aged rat brains from rat atlas (Paxinos and Watson,
920 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
Fig. 5. (A–C) Average change in MR intensity over time indicates decreased Mn2+transport to olfactory tract in aged rats. Vertical axes indicate normalized VOI
intensity subtracted from baseline. Horizontal axes indicate time post-administration in minutes. Averaged change in MR intensity of the bulb VOI over time
indicates similar magnitude across groups (A). Change in MR intensity for anterior (B) and posterior (C) tract VOIs indicates reduced Mn2+transport.
921 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
Published fast axonal transport rates from studies of peripheral
neurons indicate a range from 100 to 400 mm/day (Ochs, 1972;
Stromska and Ochs, 1982; Brunetti et al., 1987; Frolkis et al.,
1997; Verdu et al., 2000; Uchida et al., 2004). In fact, it has been
indicated that brain axonal transport rates may not be comparable
to those in the periphery (Vahlsing et al., 1981; Jacob and
O’Donoghue, 1995). The normal rate in the brain has been
reported to be 50–200 mm/day (Fibiger et al., 1972; Levin, 1977).
In comparison, our results estimate transport of manganese along
lateral olfactory tract axons to be 98 mm/day in young subjects.
This number is within the range of previously reported rates. The
large variation in normal published rates may be attributed to
different classes of molecules being transported at different rates.
Also, it has been shown that different cell types exhibit different
transport rates (Ochs, 1972). Our technique of intranasal injection
followed by interneuronal transport is easily visualized over the
entire time of experiment and the estimated rates across the
different study groups are from in the same neuronal population,
i.e. mitral cell axons, using a uniform molecule, manganese.
Several factors have been suggested that may contribute to age-
related decrease in axonal transport. Aged neurons exhibit a
decline in metabolic function, which directly impacts transport, as
the process is ATP-dependent (Frolkis et al., 1997; Verdu et al.,
2000). Of even higher potential impact on transport mechanisms,
many investigators have shown an age-related decrease in the
production of the cytoskeletal framework, such as neurofilaments
and microtubules that are critical to the functioning of neuronal
transport mechanisms (Ochs, 1973; McQuarrie et al., 1989; Parhad
et al., 1995; Cash et al., 2003; Uchida et al., 2004). Previously it
was thought that so-called “fast transport” of vesicles and synaptic
proteins was a separate process from “slow transport” of certain
cargos such as cytoskeletal components. It is now believed that
rates differences between slow and fast are a product of the amount
of time a cargo spends “associated” with the motor proteins. Where
transport is not a continuous smooth process, but is comprised of
many small steps and stops (Ochs, 1975; Shea, 2000). If the cargo
spends longer time associated with the motor and less time
pausing, than the overall rate from cell body to synapse is quicker.
The proteins involved in regulation of axonal transport are not well
understood, but we may speculate that aging and/or disease can
affect this process on multiple levels. Further development of in
vivo quantification of axonal transport will allow longitudinal
studies of interventions and drug treatments that will help elucidate
better these important processes.
In a review paper, Schwartz describes, in general, the two main
methods that have been used to quantify axonal transport as
“destination analysis” and “direct kinetic analysis” (Schwartz,
1979). Investigations using destination analysis have a site that is
injected with radioactive precursors and a destination is monitored
for the arrival of the radioactivity (Grafstein and Laureno, 1973;
Levin, 1977). Kinetic analysis usually involves injection of
radiolabeled proteins into peripheral or spinal cord neurons and
assessing counts in uniformly cut segments (Ochs, 1972; Stromska
and Ochs, 1982; Brunetti et al., 1987; Jacob and O’Donoghue,
1995; Frolkis et al., 1997; Uchida et al., 2004). More recent
investigations have studied axonal transport mechanisms using
time lapse recording of fluorescent probes in cultured cell axons
(Smith et al., 2001). Previous studies using the Mn-MRI technique
to investigate brain neuroanatomical connections have reported
axonal transport rate estimates. In one of the first papers to pioneer
this technique, Pautler et al. used intravitreal injection of MnCl2in
a mouse model and reported enhanced contrast in the optic nerve
within 2 h (2 mm/h) (Pautler et al., 1998). Another group,
Watanabe et al. performed an intravitreal injection in the rat and
saw a “weak but unequivocal enhancement of the superior
colliculus” at 8 h post administration to yield an approximate rate
Fig.6. Areaundercurve calculationsover an initialof 72 h indicatereduced bulktransportof material inaged rats.Vertical axis indicatespercent oftotal intensity
and ratios are indicated in horizontal labels. Ratios indicate the amount of material that is transported from bulb to anterior and bulb to posterior tract. Rightmost
group of columns indicates ratio of posterior to anterior tract indicating conservation of bulk material transport. Asterisk (⁎) indicates that aged rat group was
significantly lower in all three ratios than young group (p≤0.05).
922 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
of 2.8 m/h (Watanabe et al., 2001). Saleem et al. calculated rates
ranging from 0.64 to 1.42 mm/h after intracortical injections in the
caudate or putamen to globus pallidus and substantia nigra
destinations in monkeys (Saleem et al., 2002). In a study
investigating the songbird vocal system, Van der Linden et al.
estimated transport of manganese injected into the high vocal
center and traveling to the nucleus robustus archistriatalis as
ranging from 2 to 6 mm/h (Van der Linden et al., 2002). Leergaard
Fig. 7. (A–C) Time to peak intensity is delayedin mid-age and aged rats in all regions. Vertical axis indicates normalized intensity subtracted from baseline in the
olfactorybulbVOI(A), anteriorolfactorytractVOI(B), andposteriortractVOI(C).Asterisk(⁎) indicatessignificantdifference inintensitybetweenmid-agedor
aged groups and young group at the specific time point (p≤0.05).
923 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
et al. employed a similar method to calculate intracortical
manganese transport in the rat brain at 2.1–2.6 mm/h and at 4.6–
6.1 mm/h in descending corticofugal pathways (Leergaard et al.,
2003). While all of these studies reported transport rate estimates
that are within the expected range for fast axonal transport, they
employ a rough type of “destination analysis” using arbitrarily
defined changes in MR signal intensity at a predetermined time
point for the calculation. Statistical mapping and calculation of
bulk transport indicated age-related decreased transport. To support
these results, our study employed tracer kinetic estimation of
dynamic change in enhancement along a single tract over serial
scans (Fig. 1). This method allowed a more precise estimate of
Fig. 8. Time to peak flow is decreased in mid-age and aged subjects in both anterior and posterior tract. Curve-fitting parameter estimation using olfactory bulb
uptake as input to changes in anterior and posterior intensity indicates the time to peak flow in each VOI. Vertical axis indicates time. Asterisk (⁎) indicates
significant difference in time to peak flow between mid-aged or aged groups and young group for both regions (p≤0.05).
Fig. 9. Decreased neuronal transport correlates to increased age. Following time to peak flow estimation by curve-fitting, axonal transport rate between the two
tract VOI was calculated by dividing the scaled tract distance by the time difference (dist(P−A))/(time peak flow P−time peak flow A). Trendline indicates
axonal transport rates decreased with increasing age (Pearson's correlation coefficient=−0.82).
924 D.J. Cross et al. / NeuroImage 39 (2008) 915–926
individual transport rates and, since the entry into the system (the
olfactory bulb uptake) was used as the input function, calculated
transport rates can be considered “pure” with respect to manganese
transport down mitral cell axons. In comparison to previous studies
that employed direct kinetic methods, we have applied it
longitudinally in living brains.
Our study also showed an age related decrease in bulk transport
of material to the posterior tract over the initial 72 h. Age related
decrease in bulk transport was not due to olfactory tract atrophy.
Number of olfactory bulb mitral cells remains constant until rats
reach age of 24 months. In fact the cross-sectional area of the tract
may even increase slightly (Curcio et al., 1985). Therefore, the
reduced transport rate in aged rats resulted in reduced bulk
transport of material over an equal time period.
There is a potential effect of the non-linearity of MRI signal
intensity on the estimated transport rate that should be discussed.
Manganese ion concentration has been shown to have a linear
relationship (within similar tissue types) with shortening the MRI
relaxation time (T1) (Gallez et al., 2001). However this relation-
ship does not always hold true when T1-weighted imaging is
used, as in this study. Within a specific brain tissue, increasing
concentrations result in increased signal intensity on T1-weighted
images; however, lower concentrations may not be detectable in a
linear fashion due to lack of sensitivity of MR scanner. With the
issue of potential reduced sensitivity of scanner to detect small
amounts of enhancement in a linear manner, we routinely exclude
subjects that have insufficient bulb enhancement (excessive
sneezing or a nicked blood vessel occasionally result in reduced
manganese uptake in through the epithelium). A threshold of
100% peak signal increase over baseline in the bulb has been
determined empirically to yield consistent signal increases in the
olfactory tract for a given age or study group. An independent
method to assess actual Mn2+tissue concentration is under devel-
opment, but in this study, we assume that the tissue concentrations
fall within the approximate linear range of increasing intensity as
validated in our laboratory (data not shown). Since our transport
rate analysis is based on relative intensity changes within the
same tissues for each subject independently, we feel that the effect
of non-linearity is small in regards to the interpretation of the
In conclusion, our study indicated age-related decreases in
axonal transport rate and bulk transport of material in the lateral
olfactory tract of living rat brains. These results were indicated in
both longitudinal and separate group analyses. The use of kinetic
analysis combined with dynamic manganese enhanced MR
imaging provides a unique opportunity to study this crucial neuro-
nal process. To date, there have been no published studies using
Mn-MRI to estimate longitudinally cortical axonal transport rates
in aged rats. The prevalence of age-related neurodegenerative
diseases is predicted to increase over the next few decades and
studies of normal aging and pathological mechanisms are of critical
importance. Future studies will expand the use of this technique to
transgenic animal models and human disease applications (Cross
et al., 2006a).
This study is supported in part by Washington Alzheimer
Disease Research Center Pilot Research grant, Washington
National Primate Research Center grant NCRR P51 RR000166-
41, and NINDS RO1 NS045254-01.
Because of the limited transverse migration of tracers across
axons, the lateral olfactory tract can be approximated as a one-
dimensional system. Dilution and first-order decay of a pulse of the
manganese tracer in such a system can be described by an
advection–diffusion equation (Ho, Schlosser et al., 2002). For an
instantaneous injection of the tracer, the average cross-sectional
concentration of the tracer ‘c’ at a distant ‘x’ and time ‘t’ can be
where ‘u’ is the average cross-sectional transport rate, ‘Kx’ is the
due to interstitial clearance of the tracer, ‘M’ is the amount of tracer
injected, and ‘A’ is the cross-sectional area of the tract. Because of
the primary interest in this application is the rate of transport, ‘M’
and ‘A’ are combined with the relative intensity of MRI as a single
scaling factor. The equation based on an instantaneous injection of
the tracer was convolved with the olfactory bulb input function to
model continuous input of manganese from the nasal cavity during
the imaging period. Parameters were estimated by nonlinear curve
fitting to observed data.
cðx;tÞ ¼ M exp ?ðx ? utÞ2=4Kxt ? kt
=A sqrt 4pKxt
Albers, M.W., Tabert, M.H., et al., 2006. Olfactory dysfunction as a
predictor of neurodegenerative disease. Curr. Neurol. Neurosci. Rep. 6
and mild cognitive impairment. Clin. Neuropathol. 25 (6), 265–721.
Barrios, F.A., Gonzalez, L., et al., 2007. Olfaction and neurodegeneration in
HD. NeuroReport 18 (1), 73–76.
Brenneman, K.A., Wong, B.A., et al., 2000. Direct olfactory transport of
inhaled manganese ((54)MnCl(2)) to the rat. Toxicol. Appl. Pharmacol.
169 (3), 238–248.
Brunetti, M., Miscena, A., et al., 1987. Effect of aging on the rate of axonal
transport of choline–phosphoglycerides. Neurochem. Res. 12 (1), 61–65.
Cash, A.D., Aliev, G., et al., 2003. Microtubule reduction in Alzheimer’s
disease and aging is independent of tau filament formation. Am. J.
Pathol. 162 (5), 1623–1627.
Cross, D.J., Minoshima, S., et al., 2004. Statistical mapping of functional
olfactory connections of the rat brain in vivo. NeuroImage 23 (4),
Cross, D.J., Flexman, J.A., et al., 2006a. In vivo MRI of age related changes
in axonal transport in brains expressing Alzheimer pathology. Society
For Neuroscience 36th Annual meeting, Atlanta, GA, USA.
Cross, D.J., Flexman, J.A., et al., 2006b. In vivo imaging of functional
disruption, recovery and alteration in rat olfactory circuitry after lesion.
NeuroImage 32 (3), 1265–1272.
Curcio, C.A., McNelly, N.A., et al., 1985. Aging in the rat olfactory system:
relative stability of piriform cortex contrasts with changes in olfactory
bulb and olfactory epithelium. J. Comp. Neurol. 235 (4), 519–528.
Enwere, E., Shingo, T., et al., 2004. Aging results in reduced epidermal
growth factor receptor signaling, diminished olfactory neurogenesis,
and deficits in fine olfactory discrimination. J. Neurosci. 24 (38),
Fibiger, H.C., Pudritz, R.E., et al., 1972. Axonal transport in nigro-striatal
and nigro-thalamic neurons: effects of medial forebrain bundle lesions
and 6-hydroxydopamine. J. Neurochem. 19 (7), 1697–1708.
Friston, K.J., Passingham, R.E., et al., 1989. Localisation in PET images:
direct fitting of the intercommissural (AC–PC) line. J. Cereb. Blood
Flow Metab. 9 (5), 690–695.
925D.J. Cross et al. / NeuroImage 39 (2008) 915–926
Friston,K.J.,Frith,C.D.,etal.,1990.Therelationshipbetweenglobalandlocal Download full-text
changes in PETscans. J. Cereb. Blood Flow Metab. 10 (4), 458–466.
Frolkis, V.V., Tanin, S.A., et al., 1997. Age-related changes in axonal
transport. Exp. Gerontol. 32 (4–5), 441–450.
Gallez, B., Demeure, R., et al., 2001. Non invasive quantification of
manganese deposits in the rat brain by local measurement of NMR
proton T1 relaxation times. Neurotoxicology 22 (3), 387–392.
Grafstein, B., Laureno, R., 1973. Transport of radioactivity from eye to
visual cortex in the mouse. Exp. Neurol. 39 (1), 44–57.
Henriksson, J., Tjalve, H., 2000. Manganese taken up into the CNS via the
olfactory pathway in rats affects astrocytes. Toxicol. Sci. 55 (2), 392–398.
Ho, D.T., Schlosser, P., et al., 2002. Determination of longitudinal dispersion
coefficient and net advection in the tidal Hudson River with a large-
scale, high resolution SF6 tracer release experiment. Environ. Sci.
Technol. 36 (15), 3234–3241.
Itaya, S.K., 1987. Anterograde transsynaptic transport of WGA–HRP in rat
olfactory pathways. Brain Res. 409 (2), 205–214.
Jacob, J.M., O’Donoghue, D.L., 1995. Direct measurement of fast axonal
transport rates in corticospinal axons of the adult rat. Neurosci. Lett. 197
Leergaard, T.B., Bjaalie, J.G., et al., 2003. In vivo tracing of major rat brain
pathways using manganese-enhanced magnetic resonance imaging and
three-dimensional digital atlasing. NeuroImage 20 (3), 1591–1600.
Levin, B.E., 1977. Axonal transport of [3H]fucosyl glycoproteins in
noradrenergic neurons in the rat brain. Brain Res. 130 (3), 421–432.
Luzzi, S., Snowden, J.S., et al. (2006). Distinct patterns of olfactory
impairment in Alzheimer’s disease, semantic dementia, frontotemporal
dementia, and corticobasal degeneration. Neuropsychologia.
McQuarrie, I.G., Brady, S.T., et al., 1989. Retardation in the slow axonal
transport of cytoskeletal elements during maturation and aging.
Neurobiol. Aging 10 (4), 359–365.
Minoshima, S., Koeppe, R.A., et al., 1993. Integrated and automated data
analysis method for neuronal activation studies using [O-15]water PET.
In: Uemura, K., Lassen, N.A., Jones, T., Kanno, I. (Eds.), Quantification
of Brain Function, Tracer Kinetics and Image Analysis in Brain PET.
Excerpta Medica, Amsterdam, pp. 409–415.
Narita, K., Kawasaki, F., et al., 1990. Mn and Mg influxes through Ca
channels of motor nerve terminals are prevented by verapamil in frogs.
Brain Res. 510 (2), 289–295.
Ochs, S., 1972. Fast transport of materials in mammalian nerve fibers.
Science 176 (32), 252–260.
Ochs, S., 1973. Effect of maturation and aging on the rate of fast axoplasmic
transport in mammalian nerve. Prog. Brain Res. 40 (0), 349–362.
Ochs, S., 1975. A unitary concept of axoplasmic transport based on the
transport filament hypothesis. In: Bradley, W., et al. (Eds.), Recent
Parhad, I.M., Scott, J.N., et al., 1995. Axonal atrophy in aging is associated
witha decline in neurofilament geneexpression. J. Neurosci.Res.41 (3),
Pautler, R.G., Koretsky, A.P., 2002. Tracing odor-induced activation in the
olfactory bulbs of mice using manganese-enhanced magnetic resonance
imaging. NeuroImage 16 (2), 441–448.
Pautler, R.G., Silva, A.C., et al., 1998. In vivo neuronal tract tracing using
manganese-enhanced magnetic resonance imaging. Magn. Reson. Med.
40 (5), 740–748.
Paxinos, G., Watson, C., 1998. The Rat Brain in Stereotaxic Coordinates.
San Diego, Academic Press.
Pigino, G., Morfini, G., et al., 2003. Alzheimer’s presenilin 1 mutations
impair kinesin-based axonal transport. J. Neurosci. 23 (11), 4499–4508.
Saleem, K.S., Pauls, J.M., et al., 2002. Magnetic resonance imaging of
neuronal connections in the macaque monkey. Neuron 34 (5), 685–700.
Schwartz, J.H., 1979. Axonal transport: components, mechanisms, and
specificity. Annu. Rev. Neurosci. 2, 467–504.
Shea, T.B., 2000. Microtubule motors, phosphorylation and axonal transport
of neurofilaments. J. Neurocytol. 29 (11–12), 873–887.
Sloot, W.N., Gramsbergen, J.B., et al., 1994. Axonal transport of manganese
and its relevance to selective neurotoxicity in the rat basal ganglia. Brain
Res. 657 (1–2), 124–132.
Smith, G.A., Gross, S.P., et al., 2001. Herpesviruses use bidirectional fast-
axonal transport to spread in sensory neurons. Proc. Natl. Acad. Sci.
U. S. A. 98 (6), 3466–3470.
Stokin, G.B., Goldstein, L.S., 2006a. Axonal transport and Alzheimer’s
disease. Annu. Rev. Biochem. 75, 607–627.
Stokin, G.B., Goldstein, L.S., 2006b. Linking molecular motors to
Alzheimer’s disease. J. Physiol. (Paris) 99 (2–3), 193–200.
Stromska, D.P., Ochs, S., 1982. Axoplasmic transport in aged rats. Exp.
Neurol. 77 (1), 215–224.
Takeda, A., Kodama, Y., et al., 1998. Manganese transport in the neural
circuit of rat CNS. Brain Res. Bull. 45 (2), 149–152.
Tjalve, H., Henriksson, J., 1999. Uptake of metals in the brain via olfactory
pathways. Neurotoxicology 20 (2–3), 181–195.
Uchida, A., Tashiro, T., et al., 2004. Morphological and biochemical
changes of neurofilaments in aged rat sciatic nerve axons. J. Neurochem.
88 (3), 735–745.
Vahlsing, H.L., Hirschl, R.B., et al., 1981. Axoplasmic flow of tritiated
proline in the corticospinal tract of the rat. Cell Tissue Res. 214 (2),
Van der Linden, A., Verhoye, M., et al., 2002. In vivo manganese-enhanced
magnetic resonance imaging reveals connections and functional pro-
perties of the songbird vocal control system. Neuroscience 112 (2),
Verdu, E., Ceballos, D., et al., 2000. Influence of aging on peripheral nerve
function and regeneration. J. Peripher. Nerv. Syst. 5 (4), 191–208.
Watanabe, T., Michaelis, T., et al., 2001. Mapping of retinal projections in
the living rat using high-resolution 3D gradient-echo MRI with Mn2+-
induced contrast. Magn. Reson. Med. 46 (3), 424–429.
Worsley, K.J., Evans, A.C., et al., 1992. A three-dimensional statistical
analysis for CBF activation studies in human brain [see comments].
J. Cereb. Blood Flow Metab. 12 (6), 900–918.
Worsley, K.J., Marrett, S., et al., 1996. A unified statistical approach for
determining significant signals in location and scale space images of
cerebral activation. In: Myers, R., Cunningham, V., Bailey, D., Jones, T.
(Eds.), Quantification of Brain Function Using PET. Academic Press,
San Diego, pp. 327–333.
926D.J. Cross et al. / NeuroImage 39 (2008) 915–926