|Home | About | Journals | Submit | Contact Us | Français|
Connections from hippocampus to septal nuclei have been implicated in memory loss and the cognitive impairment in Down syndrome (DS). We trace these connections in living mice by Mn2+ enhanced 3D MRI and compare normal with a trisomic mouse model of DS, Ts65Dn. After injection of 4 nl of 200 mM Mn2+ into the right hippocampus, Mn2+ enhanced circuitry was imaged at 0.5, 6, and 24 hr in each of 13 different mice by high resolution MRI to detect dynamic changes in signal over time. The pattern of Mn2+ enhanced signal in vivo correlated with the histologic pattern in fixed brains of co-injected 3kD rhodamine-dextranamine, a classic tracer. Statistical parametric mapping comparing intensity changes between different time points revealed that the dynamics of Mn2+ transport in this pathway were surprisingly more robust in DS mice than in littermate controls, with statistically significant intensity changes in DS appearing at earlier time points along expected pathways. This supports reciprocal alterations of transport in the hippocampal-forebrain circuit as being implicated in DS and argues against a general failure of transport. This is the first examination of in vivo transport dynamics in this pathway and the first report of elevated transport in DS.
Down syndrome, trisomy 21, is the most frequent genetic cause of mental retardation (Epstein, 1986, Korenberg et al., 1994). A mouse model of Down syndrome (DS), the partial trisomic Ts65Dn, has been extensively characterized with respect to deficits in behavioral measures for learning and memory, histopathology of the brain, and functional parameters (Davisson et al., 1990, Davisson et al., 1993, Reeves et al., 1995, Olson et al., 2004, Roper and Reeves, 2006, Seregaza et al., 2006). As in human DS, Ts65Dn mice display decreased cholinergic neuron density in the medial septal nucleus (MSN) of the forebrain (Holtzman et al., 1996, Cooper et al., 2001, Salehi et al., 2003, Delcroix et al., 2004), as well as over-expression of genes encoded on the excess chromosomal segment (Holtzman et al., 1996, Hunter et al., 2003, Seo and Isacson, 2005).
Reciprocal connections link neurons in the forebrain with those in the hippocampus. Neurons with cell bodies in the forebrain, including glutaminergic, GABAergic and cholinergic neurons, send processes to the hippocampus (Amaral and Kurz, 1985, Gaykema et al., 1991, Alonso et al., 1996, Risold and Swanson, 1996, Risold and Swanson, 1997, Colom, 2006) and hippocampal pyramidal neurons innervate the septum (Wu et al., 2000). Transport within these processes is thought to be important for neuronal functioning and survival (Stokin et al., 2005). When radiolabeled nerve growth factor (NGF) is injected into the hippocampal region, less arrives in the forebrain in Ts65Dn mice compared to wildtype littermates (Cooper et al., 2001). This has been attributed to decreased retrograde transport in TrkA expressing cholinergic neurons and thus is proposed to underlie the cholinergic neuronal atrophy in these mice. Treatment with NGF by intracerebroventricular injection in Ts65Dn mice, nonhuman primates, and Alzheimer's patients results in recovery of basal forebrain cholinergic cells expressing the NGF receptor, TrkA (Tuszynski et al., 1990, Cooper et al., 2001, Tuszynski et al., 2005). Thus, decreased transport of NGF from hippocampus to forebrain may play a significant role in DS, although direct imaging of such transport is lacking. Recent work also implicates transport in this pathway as a causative factor in Alzheimer's disease, where axonal varicosities presumed to represent transport blockages were found in cholinergic neurons projecting from septum to hippocampus (Stokin et al., 2005). Taken together these results suggest that generalized defects in transport may result in cognitive dysfunction and neuronal degeneration.
Here we test whether connections from the hippocampus to the forebrain are altered in DS mice. We use Mn2+ enhanced magnetic resonance imaging (MEMRI) to image this circuit in living DS mice and their normal 2N littermates. Mn2+ is a T1 MR contrast agent that enters neurons through calcium channels (Merritt et al., 1989), is carried down the axon primarily by anterograde transport (probably via transport vesicles), and crosses synapses (Lin and Koretsky, 1997, Pautler et al., 2003, Silva et al., 2004). In the macaque, stereotaxic co-injection of Mn2+ and the classic histologic tracer, wheat germ agglutinin conjugated to horse radish peroxidase, yielded the same highly specific projection pattern (Saleem et al., 2002, Murayama et al., 2006), further demonstrating the usefulness of Mn2+ as a trans-synaptic tracer of neuronal connections.
We injected nanoliter volumes of Mn2+ into a precise location in the right hippocampus of 7 Ts65Dn and 6 2N littermates and imaged the living mice by MR at 0.5, 6 and 24 hr afterwards. To obtain statistically reliable information, we performed 3-dimensional statistical analyses of Mn2+-induced changes of intensity over time. First images obtained at the same time point from all mice in either the Ts65Dn or 2N littermates group were warped into one 3D space (Lancaster et al., 2003, Kovacevic et al., 2005, Lee et al., 2005). By taking the average intensity of each voxel from all warped images in each group a single averaged 3D image for each genotype at each time point was generated. Datasets from two different time points were compared using the paired Student's t-test to identify those voxels whose intensity changed significantly between time points. Co-injection of the classical histological tracer, 3kD rhodamine-dextran-amine (RDA), followed by post-mortem microscopic analysis confirmed our MR methodology. In contrast to studies showing decreased NGF transport in DS mice, we find robust transport in these mice of these general tract tracers. Similarities in the molecular machinery used in anterograde and retrograde transport allows us to draw conclusions about general transport dysfunction in the DS mouse.
Nine to eleven month old Ts65Dn mice and 2N littermates were obtained from Jackson Labs (Bar Harbor, ME). Seven Ts65Dn (trisomic) and six 2N littermates (wildtype) mice were employed in this study. All mice were male. Genotypes were confirmed by Cecilia Schmidt (Jackson Labs) by quantitative PCR (Liu et al., 2003). All animal experiments conformed to National Institutes of Health guidelines and were approved by the California Institute of Technology Animal Care and Use Committee.
Calibrated micropipettes were used to attain consistent injection volumes. The micropipettes were fashioned from 1 mm OD quartz capillary pulled to approximately 40μm ID bore near the tip with a Sutter Instruments (Novato, CA) P-2000 micropipette puller. The identical program was used to pull all micropipettes. Micropipettes were calibrated by drawing a 4mm length of the MnCl2/RDA solution (see below) into the end of the micropipette using a World Precision Instruments Inc. (Sarasota, Florida) microsyringe pump with digital display driving a 10μl syringe fitted to the micropipette via oil-filled 0.01”ID PEEK™ (polyetheretherketone) tubing (Upchurch Scientific Inc, Oak Harbor, WA). The total solution volume was slowly ejected into a small amount of oil on a hemocytometer using the microsyringe pump, and the volume of the spherical drop determined from measurement of its diameter. Performing this procedure with six different micropipettes yielded a volume of 4.3±0.6 nl (average ± standard deviation) for the delivered material. The same apparatus and filling/ejection procedure was used for stereotaxic injections.
Ts65Dn and 2N littermate mice were anesthetized with ketamine/xylazine (7.5mg ketamine plus 5mg xylazine per kg, i.p) and placed in a stereotaxic frame (Kopf Instruments, Tujunga, CA). 4.3 nl of 200 mM MnCl2 with 0.5 mg/ml RDA (3k) (Molecular Probes/Invitrogen, Eugene, OR) was injected unilaterally into the right hippocampus (coordinates x –3.2 mm (midline), y -4.1 mm (Bregma), z 3.4 mm (down) (Paxinos and Franklin, 2001)) over 5 minutes using the microsyringe pump and calibrated micropipette in a modified procedure (Bearer et al., 2000, Jaffe and Terasaki, 2004). After injection, the capillary was left in place for 3-5 minutes and withdrawn over 5 minutes. Animals were injected and imaged over a 3 week time period in random order.
Following injection, the animal was immediately placed in the MR scanner under 0.8% isoflurane anesthetic. Inspection of MR images recorded 0.5 hr after injection revealed that the average injection site (center of the small hypointense region) for the Ts65Dn mice was: x (lateral) +3.43±0.37 mm; y (A-P), -3.55±0.4 mm, z (D-V) -4.07±1.5 mm; and for 2N littermates was: x, +3.27±0.22 mm; y, -4.23±0.23 mm; z, -3.05±0.26 mm.
An 11.7T 89 mm vertical bore Bruker BioSpin Avance DRX500 scanner (Bruker BioSpin Inc, Billerica, MA) equipped with a Micro2.5 gradient system was used to acquire in vivo mouse brain images with a 35 mm linear birdcage RF coil. The animal's head was secured in a Teflon stereotaxic unit within the RF coil to minimize movement and aid in reproducible placement. Temperature and respiration were continuously monitored during data acquisition and remained within normal ranges. We employed a 3D RARE imaging sequence (Hennig et al., 1986) with RARE factor of 4, 4 averages, TR/TEeff = 300ms/21ms; matrix size of 256×160×128; FOV 23 mm × 14.4 mm × 11.5 mm; yielding 90 μm isotropic voxels with 102 minutes scan time. Images were recorded from each mouse before (1-5 days) and after injection (0.5 hr, 6 hr, and 24 hr). All 13 mice (7 Ts65Dn and 6 2N littermates) were imaged at each of these 4 time points.
One day to 3 weeks after in vivo MR imaging, each mouse was anesthetized, sacrificed and fixed by transcardial perfusion with a 30 ml washout with warm heparinized phosphate buffered saline (PBS) followed by 30 ml of room temperature 4% paraformaldehyde (PFA) in PBS at a rate of 5 ml/min. The carcass was decapitated and the head rocked in 4% PFA in PBS overnight at 4°C. Brains dissected from the calvarium were sent to Neuroscience Associates (NSA, Knoxville, TN) for gelatin embedding, frozen sectioning at 35 μm and staining. Alternate sections were selected for staining with either Thionine/Nissl for cellular morphology, or immunohistochemistry for anti-choline acetyl transferase (ChAT), a relatively specific marker for cholinergic neurons, or for mounting unstained in anti-quench (Vector Lab, Burlingame, CA) to image the RDA fluorescence. Sections were imaged on a Zeiss Axioscope equipped with 5×, 10×, 20×, 40×, 63× and 100× neofluor objectives and captured by an Axoicam MRM. The number of ChAT positive neurons in the MSN was obtained from three trisomic and three 2N littermates embedded in the same gelatin block and sectioned in parallel simultaneously in register. Four sections traversing the MSN were selected for counts, and all six brains were counted in each section. The anterior commissure was used as a landmark to confirm equivalency of the position of the section for each brain along the A-P axis. More than 1,000 cells were counted for each animal.
MR image data was converted to Analyze 7.5 format (Mayo Clinic, Rochester, MN) modified to be readable in SPM2 (Wellcome Department of Imaging Neuroscience, University College London). SPM2 was used for all statistical comparisons in a manner similar to Cross and coworkers (Cross et al., 2004). After bias correction in SPM2, each was scaled to the mode of its intensity histogram (Kovacevic et al., 2005). Initially, one of the wildtype pre-injection images was chosen as a template and all 52 images (all animals at all time points) were normalized to it using an affine transformation and interpolated to 50μm voxel size. The average of the normalized pre-injected images from the wildtype group and from the DS group were each calculated and used as template images for their respective groups; removing bias due to selection of any individual image as template (Kovacevic et al., 2005, Lee et al., 2005). Pre-injected DS and 2N brain images were then normalized to their respective averages using full nonlinear normalization. The nonlinear transformation used to drive each individual mouse's pre-injected scan into its template was re-applied to drive that mouse's 0.5, 6, and 24 hr scans into the same coordinate system. Images were smoothed to 300μm and paired Student's t-tests performed comparing the different time points within each group, providing t-value and corresponding p-value maps of significant differences. The averaged images shown in Figures 3, ,55 and and66 represent the arithmetic mean of each voxel in the smoothed DS or 3N littermate individual images.
Visualization of the MR images and statistical parametric maps was performed with ResolveRT4 (Mercury Computer Systems, Inc., Hudson, NH) and MRIcro (Rorden and Brett, 2000).
For each of the 13 half hour data sets, the number of voxels within injection site halo were determined using the 3D ROI feature in MRIcro. The halo was defined as contiguous voxels around the injection site with intensity greater than the mean unenhanced parenchyma intensity plus 3 times the background noise level.
The 3D maps of voxels with significantly increased intensity between each pair of images (0.5 hr versus preinjection, 6 versus 0.5 hr, and 24 versus 6 hr) were used to determine the volume of brain displaying Mn2+ enhanced signal in each group. The number of voxels within a defined anatomical region whose intensity was enhanced (p<0.0005) was computed using MRIcro.
The Ts65Dn genotype was confirmed by PCR and the phenotype by histologic examination of individual mice fixed by perfusion after MRI (Fig. 1). Ts65Dn mice displayed the expected morphological phenotype of decreased cell numbers and loss of cholinergic neurons in the septal nuclei (Fig. 1), as previously reported (Granholm et al., 2000, Cooper et al., 2001, Seo and Isacson, 2005). Histological examination revealed qualitative decreases in the number and size of cells in the septal nuclei of DS mice as compared to 2N littermates (Fig. 1A and B). Immunohistochemistry with anti-choline acetyl transferase (ChAT) demonstrated diminished numbers of cholinergic neurons in the medial septal nuclei (MSN) of DS mice versus 2N littermates (Fig. 1C). In six of the mice, counts of ChAT positive nuclei in histologic sections spanning the MSN demonstrated a 30 ± 5% decrease in the number of neurons in DS versus 2N littermates. ChAT-positive neuronal cell bodies found in the MSN of DS brains also appeared shrunken compared to the plump cell bodies in littermates. No obvious differences in thickness or number of fibers in the hippocampal commissure, fimbria, or fornix were observed in histological sections. Thus, the aging Ts65Dn mice used in this study displayed the expected morphological abnormalities previously described for these mice, features similar to those found in human Alzheimer's disease and Down syndrome.
Co-injection of the tracer, RDA, with the Mn2+ allowed us to correlate the injection site in histologic sections with MR images of the same individuals. Imaging by combined phase and fluorescent microscopy of unstained gelatin-embedded sections of brains fixed immediately after acquisition of the last MR image (24 hr) identified the precise location of the injection (Fig. 2). In animals fixed 6 days after injection, no needle track or tissue damage at the injection sites were identified.
Comparison of MR images captured 0.5 hr after injection demonstrated that the injection position was accurately and reproducibly placed in all the mice studied (Fig. 3). The epicenter of the injection site (±200μm diameter), containing the highest concentration of Mn2+, appears hypointense (dark) in MR images at this early time point because the level of Mn2+ there was sufficient to produce significant T2-shortening and concomitant darkening (Lee, 1991). Analysis of the 3D location of this epicenter in each of the 13 animals demonstrated that among all animals injections were routinely placed in CA3 of the hippocampus (see Materials and Methods).
At 0.5 hr post injection, the hypointense injection site was surrounded by a white halo of hyperintensity indicating passive diffusion of the Mn2+ ion outward from the injection site. The halo volume was found to be 9.3±1.9 μl for the Ts65Dn cohort and 9.0±1.4 μl for the 2N littermates. An equivalent spherical volume would have a radius of 1.3 mm. Such diffusion is a common factor in all studies employing water-soluble low molecular weight tracers. Previous studies in the songbird, mouse, rat, minipig, and macaque note the same initial Mn2+ diffusion with no lack of specificity of subsequent projection patterns (Saleem et al., 2002, Van der Linden et al., 2002, Leergaard et al., 2003, Pautler et al., 2003, Van der Linden et al., 2004, Watanabe et al., 2004b, Jelsing et al., 2006, Murayama et al., 2006). Presumably, only neurons in the immediate vicinity of the high Mn2+ concentration take up enough Mn2+ for detection after transport to distant sites. To focus the high concentration of Mn2+ in as small a brain region as possible, the volume of Mn2+ injected in this study (4.3 nl) was 10-1000 times less than that employed in previous studies. The reproducible placement of the injection sites and the comparable size of the diffusion halo argue that consistent amounts of both tracers, Mn2+ and RDA, were delivered to the same population of neurons in each animal tested. Thus, comparisons between individual animals or across cohorts are not confounded by effects due to different amounts of tracer introduced or different locations being probed.
Co-injection of RDA with the Mn2+ allowed tracing of fluorescence in histologic sections of the same animals imaged by MR (Fig. 4). By comparing the patterns of distribution of Mn2+ and RDA, we confirmed that the placement of the injection site was appropriate for entry of tracer into the hippocampal-septal connections, and that the transport within those connections was operational. Small RDA such as the 3kD used here are taken up and moved inside neuronal processes, either diffusively or primarily in the retrograde direction at slow rates. Diffusive motion of small dextrans is observed in the squid axon (Terasaki et al., 1995) and 1-3 weeks is required for studies of CNS pathways in adult rodents (Reiner et al., 2000, Van Haeften and Wouterlood, 2000). In serial sections of brains fixed 6 days after injection, RDA was found along the fimbria and fornix ipsilateral to the injection site but not detectable on the contralateral, non-injected, side (Fig. 4A-C). At the level of the medial septal nucleus, containing the cholinergic neurons that send processes to the hippocampus, RDA fluorescence appeared on both sides of the midline (Fig. 4D).
These results confirm that the injection site in the hippocampus was positioned appropriately for entry into the hippocampal-septal pathway. No other locus with detectible accumulation of RDA was found. These results did not reveal any obvious difference in RDA fluorescence between DS mice and their 2N littermates in the fimbria, fornix or septal region. This is consistent with reports of another retrograde tracer, Fluorogold, which was injected in CA3 of the hippocampus in much higher quantities than in our study, and arrived into the MSN of Ts65Dn and their littermates in similar amounts (Salehi et al., 2006).
Qualitative assessment of MR slices taken at the level of the basal forebrain revealed a time-dependent accumulation of Mn2+ in the septal nuclei after injection into the hippocampus (Fig. 5). These coronal slices were selected from 3D MR images at Bregma +0.4 mm position (see Fig. 5A for diagram). In these averaged 3D images at 0.5 hr post-injection, no Mn2+ enhanced brightness was seen in the septal region although slices through the injection site at Bregma -4 .0 mm display Mn2+ enhancement as shown in Fig. 3. In contrast, by 6 hr the septal region ipsilateral to the injection site appeared brighter and this persisted in the 24 hr image (Fig 5B). Mn2+ enhancement in the septal region at 6 and 24 hr is qualitatively similar or greater in Ts65n mice than in 2N littermates. Mn2+ enhanced signal also appeared in the contralateral hippocampus at 24 hr as shown in coronal slices at the level of the dentate gyrus of the hippocampus, Bregma -2.7 (Fig. 6). Again this enhancement was qualitatively similar or greater in Ts65Dn than in 2N littermates.
Thus, Mn2+ was transported from the injection site in the hippocampus to the septum in relatively short times and to the contralateral hippocampus within 24 hours. Transport from the injection site to the septum required more than 0.5 hr and less than 6 hr to traverse the distance of ~6.5 mm (from injection site, along the fimbria, to septum); we thus calculate a transport rate from 0.3 to 3.6 μm/sec, consistent with fast axonal transport of 0.05-2 μm/sec (Satpute-Krishnan et al., 2006).
SPM along with spatial realignment and normalization have become standard methodologies for human brain imaging studies (Thompson and Toga, 1996, Thompson et al., 1997, Woods et al., 1998, Hajnal et al., 2001, Toga and Thompson, 2001, Toga and Mazziotta, 2002, Lancaster et al., 2003, Yoo, 2004). The same general procedures employed here have been used to make ‘standard atlas’ spaces for mouse brains of various strains and ages (MacKenzie-Graham et al., 2004, Kovacevic et al., 2005, Lee et al., 2005). Despite significant anatomical abnormalities (Baxter et al., 2000; Kurt et al., 2000; Olson et al., 2004), the brains of the Ts65Dn mice were surprisingly similar structurally, and could be aligned into the same 3D space. However, these Ts65Dn abnormalities precluded alignment of the mutant brains into the same template as their anatomically normal littermate controls. We therefore pursued analysis of the circuits enhanced by Mn2+ within each cohort and compared the resulting patterns between groups.
To identify dynamic changes in the anatomical pattern of Mn2+ enhancement over time after injection, we used SPM to compare each subsequent dataset with that immediately preceding it in the time series. Here we present three comparisons for each genotype: 0.5 hr post-injection versus pre-injection (Fig. 7A), 6 hr versus 0.5 hr (Fig. 7B), and 24 versus 6 hr post-injection (Fig. 7C). Those voxels with significantly (P<0.0005) increased signal intensity in the later time point image are displayed as a colored voxel on the maps (Fig. 7).
In the 0.5 hr versus pre-injection comparisons (Fig. 7A), both sets of animals displayed similar paired Student's t-test maps at P<0.0005. The injection site was in a similar position and of a similar size in the posterior hippocampus in both cohorts. This result confirms the precise position of the injection site and that the amount of Mn2+ injected was similar in all animals. Relative lack of enhanced signal (colored voxels in Fig. 7A) outside the injection site attests to the dependence of signal enhancement (at P<0.0005) on the presence of Mn2+ and to the accuracy of the alignment.
Comparison of 6 hr versus 0.5 hr statistical maps demonstrated that Mn2+ followed the expected route from the hippocampus through the fimbria to the basal forebrain in both DS and 2N littermates (Fig. 7B). Voxels with statistically significant increased intensity were seen in both normal and Ts65Dn cohorts in CA3, the fimbria (fi), the septofimbrial nucleus (Sfi), and the LSN (Table 1). In the DS cohort more statistically significant voxels were found in these structures and significant signal had progressed further along the expected pathways, appearing already at 6 hours in the MSN, contralateral fimbria, as well as secondary olfactory structures (dorsal portion of acumbens core, anterior olfactory nucleus, and dorsal tenia tecta; see Table 1).
Continued transport in these living mice during the 6 hr to 24 hr interval was revealed by SPM comparisons of these two time points. In this case, only voxels whose intensity increased between 6 hr and 24 hr are displayed (Fig. 7C). This comparison revealed that, by 24 hr in 2N littermates, Mn2+ transport caught up to that seen at earlier time points in DS mice. Thus by 24 hr Mn2+ had progressed into the MSN in both sets of animals. By this time, both cohorts also displayed voxels with significantly increased intensity in CA3 of the contralateral hippocampus, and bed of nucleus stria terminalis (Table 1).
In addition, at 24 hr post-injection, the DS group again out-performed their 2N littermates, displaying Mn2+ signal further rostrally along the hippocampal-forebrain pathway, with statistically significant signal enhancement in the nucleus accumbens and areas within the ventral pallidum; the ventromedial hypothalamus and the amygdala; and the anterior part of retrosplenal granular cortex (Fig. 7C and Table 1). Thus DS mice compared to 2N littermates displayed voxels with statistically significant Mn2+ enhancement further along the hippocampal-forebrain pathway at both 6 and 24 hr after injection.
Comparison of the total volume occupied by significantly enhanced voxels (P<0.0005) in the statistical parametric maps of each set of animals demonstrated that DS had a larger volume enhanced by Mn2+ than their 2N littermates (Fig. 8). In the map comparing the 6 hr to 0.5 hr data (Fig. 7B), the volume of Mn2+ enhanced voxels in the basal forebrain occupied 2.4 mm3 in DS animals, but only 1.3 mm3 in 2N littermates. Mn2+ enhanced volumes at 24 hr in the contralateral hippocampus were 3.1 mm3 and 0.2 mm3 in the DS and 2N littermates, respectively. Thus, the extent of Mn2+ transport is greater (observed further along expected tracts) and more robust (more volume affected) in DS mice as compared to 2N littermates.
Here we show that Mn2+ injected into CA3 of the hippocampus dynamically traces the hippocampal to septum pathway during the 24 hours after injection. Transport of Mn2+ from hippocampus to basal forebrain is more robust in a trisomic mouse model of Down syndrome, Ts65Dn, than in 2N littermates. The transport pattern of Mn2+enhancement measured in live animals coincides with the pattern of a coinjected traditional tracer (RDA) measured in histological sections. Histological examination revealed the typical DS phenotype of the DS mice and little to no tissue damage by the injection was observed. Alignment and averaging of MR images demonstrated the reliability of location and amount of injectate. Voxel-wise paired Student's t-test analysis of co-registered brain images from 7 individual DS mice at different times after injection revealed an unbiased comprehensive view of the time-course and pattern of Mn2+ transport. Comparisons of SPM analyses of DS mice with 2N littermates revealed that Mn2+ induced signal enhancement occupied more voxels and progressed more quickly along expected pathways in Ts65Dn mice than in normal 2N littermates.
Mn2+, a calcium analogue, enters neurons and other cells through calcium and/or other divalent ion channels (Merritt et al., 1989, Kwan and Putney, 1990, Fasolato et al., 1993, McColl and Naccache, 1997) and is transported along neuronal pathways, even crossing synapses (Lin and Koretsky, 1997, Pautler et al., 1999, Duong et al., 2000, Leergaard et al., 2003, Watanabe et al., 2004a). Transport of Mn2+ is microtubule-dependent and primarily in the anterograde direction (Sloot and Gramsbergen, 1994, Tjalve et al., 1995, Pautler et al., 1998) at speeds consistent with fast axonal transport, suggesting that once in the cytoplasm, Mn2+ enters membrane-bound organelles and is trafficked by normal vesicular transport mechanisms. Mn2+ tract tracing has been used to reveal changes in neuronal circuitry in the songbird in response to endocrine stimulation (Van der Linden et al., 2002, Van der Linden et al., 2004).
The degree of Mn2+ induced hyperintensity at a location distant from its initial placement is dependent on at least three factors: 1) local uptake; 2) directed movement along/within the neuronal connections (including initial packaging in the cytoplasm, microtubule-based movement within processes, and possibly crossing synapses); and 3) accumulation at the distant site. Operationally we use the term “transport” to encompass all three factors. Thus, changes in transport (as evidenced by a greater or lesser amount of tracer found at the distant site) can be due to modulation of any or all of these factors. Similar caveats pertain to all tract tracers (e.g. radioactive or fluorescent) used to study neuronal projections. In addition to transport defects, DS mice may also have synaptic alterations that could effect Mn2+ uptake (Schuchmann et al., 1998, Belichenko et al., 2004).
The hippocampal to forebrain pathway contains neuronal processes emanating from either end. In each type of neuron, both anterograde and retrograde transport carry endogenous cargo to and from the cell bodies. Introduction of a traceable compound at one end of the circuit results in its transport to the other end by either anterograde or retrograde intraneuronal trafficking. Transport of NGF from the hippocampus to the basal forebrain is decreased in Down syndrome mice of the same age as those used in this work (Cooper et al., 2001, Salehi et al., 2003). This decrease is thought to be within cholinergic neurons whose cell bodies are in the MSN. Thus, decreased delivery of NGF from an injection site in the hippocampus to the septum likely reflects defects in local uptake and/or retrograde transport in these neurons. Our results suggest that this decrease may be selective for NGF transport and not a result of generalized transport deficits, since Mn2+ transport is at least as robust as in normal littermates. In support of this, retrograde transport of Fluorogold, also a general tracer for retrograde connections, is also apparently normal in these mice, although more difficult to quantify precisely (Salehi et al., 2006).
MR imaging provided three advantages over traditional approaches to the study of transport in the hippocampal to forebrain pathway. First, live animals could be imaged repeatedly, thereby obtaining a sequential view of the process of transport along the pathway. Second, three dimensional information is obtained in a holistic image acquired in one session, avoiding the artifacts of fixation, embedding and sectioning and speeding data collection. Third, quantitative analysis of pooled samples allows statistical significance across a cohort of individuals to be obtained. Thus, more reliable information from living animals in three dimensions with statistical power is possible with MRI.
SPM provides a comprehensive anatomically unbiased view of the time dependence of Mn2+ transport in the DS and normal mice. This approach revealed that the entire pathway from the hippocampus to the basal forebrain was enhanced at 6 hr compared to 0.5 hr post-injection. The volume of enhancement in the basal forebrain was almost a factor of two larger in trisomic versus normal animals; directly confirming the more robust character of transport in this Down syndrome model. At 24 hr compared to 6 hr post-injection, the volume of enhancement in the contralateral hippocampus is a factor of 12 larger in DS mice than normal animals. Thus, these observations depict an overall increase in transport efficacy in the DS mice, which also extends beyond the hippocampal to basal forebrain circuit to more distal locations.
Evidence is accumulating that axonal transport plays a significant role in neuronal survival and neurodegenerative diseases (Mandelkow et al., 2004, Morfini et al., 2005, Stokin et al., 2005, Chevalier-Larsen and Holzbaur, 2006). Four mechanisms underlying altered transport have been proposed (Goldstein, 2003, Salehi et al., 2004, Morfini et al., 2005): (1) blockage of the path, (2) abnormal tracks, (3) defects in motor machinery, and (4) disrupted signaling. Physical blockage of the axon lumen would hamper transport in both directions, as would disruption of microtubule tracks or general motor defects. We thus exclude these three possibilities as mechanisms to explain decreased transport of NGF in these mice, since we observed more robust transport of Mn2+ in the DS mouse model, while retrograde transport of low molecular weight RDA and Fluorogold transport both appear unchanged (Salehi et al., 2006). Secondly, both anterograde and retrograde motors (kinesin and dynein) use the same microtubule track system, thus a general destabilization of tracks is inconsistent with current observations concerning transport in the DS model. Our results do not rule out a selective defect in NGF transport that may involve a defect in uptake, binding/linkage of NGF-bearing cargo to the motor or signals that activate its transport.
Another plausible explanation for both more robust anterograde and selectively decreased retrograde transport in DS mice could be that high APP levels increase affinity of transport vesicles for anterograde motors at the expense of retrograde motors, thus altering the balance between anterograde and retrograde transport (Salehi et al., 2006). Among the genes that are triplicated in human Down syndrome is amyloid precursor protein (APP), a component of senile plaques of Alzheimer's disease and recently implicated as a vesicular receptor for axonal motor machinery (Satpute-Krishnan et al., 2006, Kamal et al., 2000, Kamal, 2001, Lazarov et al., 2005). APP is among the genes that co-segregates with mental retardation in partial trisomies in human (Korenberg et al., 1994). Ts65Dn mice also carry three copies of the APP gene and over-express the protein (Salehi et al., 2006, Seregaza et al., 2006). The necessary coordination of anterograde and retrograde transport (Martin et al., 1999, Gross et al., 2002, Welte, 2004, Klumpp and Lipowsky, 2005) implies that effective movement of cargo (e.g. NGF) could be disrupted without altering the detailed molecular mechanisms involved in movement along tracks, but simply by disrupting the interaction(s) that coordinate directionality.
The use of Mn2+ as a transport tracer and longitudinal magnetic resonance imaging allowed us to measure transport dynamics in vivo in individual animals. This revealed functional differences in the hippocampal to forebrain pathway in a mouse model of Down syndrome. Co-registration of images from a cohort of animals allows application of statistical methods for unbiased evaluation of spatio-temporal changes in Mn2+ transport, revealing more robust transport in the Down syndrome mouse. No other technology offers the ability to peer into the living mammalian brain and determine physiological transport parameters. These results point out that neuronal transport is a complicated phenomenon and that alterations in transport dynamics can have important ramifications in cognitive functioning. Future application of this powerful technology to earlier stages in DS and to other types of neuropathologies where transport defects are implicated will undoubtedly provide more mechanistic insights.
We thank Muriel Davisson and Cecilia Schmidt for their advice and help in obtaining and genotyping aged Ts65Dn mice and 2N littermates, to J.D. DelCroix for advice on hippocampal injections, and to the Moore Foundation for the Moore Distinguished Scholar award that brought E.L.B. to Caltech to collaborate in this work. The project was in part funded by NINDS NS046810, NIGMS GM47368 (E.L.B.), NCRR U24 RR021760 entitled Mouse BIRN and NIDA R01DA018184 (R.E.J.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.