Background
Accumulation of intracellular neurofibrillary tangles (NFTs) consisting of microtubule-associated protein tau is a major hallmark of Alzheimer’s disease (AD) and related neurodegenerative diseases regarded as ‘tauopathies’ [
1‐
3]. Findings of tau mutations in subjects affected by frontotemporal dementia and mutant tau expression systems, which lead to tau-positive inclusions, neuron loss and behavioral abnormalities in various animal models, have established a role of this protein in neurodegeneration [
4‐
11].
The rTg4510 mouse line was specifically developed to model aspects of human tauopathy, with overexpression of human tau containing the P301L mutation that is associated with frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17-Tau). Expression of human tau in the rTg4510 mouse is controlled by the tetracycline transactivator (tTA) transgene under the Ca
2+/calmodulin-dependent protein kinase IIα (CaMKIIα) promoter. This leads to selective tau expression from an otherwise transcriptionally inactive tau transgene [
12]. Mice develop robust intracellular deposition of tau protein in cortico-limbic areas, which is physiologically relevant to AD and other tauopathies, and they also show age-related forebrain atrophy.
The main pathological features of rTg4510 mice have been obtained as a result of postmortem analysis of brain tissue, except recent work employing multiphoton microscopical analysis [
13,
14]. The advent of rodent magnetic resonance imaging methods can fill the need for investigating
in vivo correlates of early signs of disease. Recent developments in the animal imaging have led to the use of the manganese ion (Mn
2+) as a neuronal contrast agent that provides a useful tool for functional brain mapping at good spatial resolution [
15,
16]. MEMRI thus provides a noninvasive approach for mapping neural activity in transgenic mice. Mn
2+ is highly paramagnetic and enhances signal intensity in T
1 weighted brain images [
17,
18]. It enters neurons largely through voltage dependent calcium channels, and possibly also via vesicular reuptake mechanisms, which correlates well with changes in synaptic firing of neuronal populations [
19‐
21]. Other MEMRI studies in rodents have shown Mn
2+ signal intensity in various ROI, including the hippocampus, under basal conditions [
17,
18]. Here, we performed
in vivo brain MEMRI of rTg4510 mice and non-transgenic (nonTg) littermates. Mutant rTg4510 mice sustain significant tau-associated hippocampal atrophy at 6 months of age and above and this was anticipated to result in less Mn
2+ signal within this region in basal un-stimulated conditions. We observed clear differences in basal neural activity between mutant tau expressing mice and controls, with the magnitude of activity being lower in rTg4510 mice. Our study extends previous applications of MEMRI in the study of axonal transport rates in transgenic mice by showing reduced neural activity in two structures involved in learning and memory [
22‐
24].
Discussion
rTg4510 mice develop NFTs and show significant loss of hippocampal neurons at around 5–6 months. Neuronal loss may be severe (~80%) by 8 months [
26]. These factors are likely to contribute to impairments observed in spatial navigation memory previously reported using the water maze test [
27], a behavior that is associated with neuronal activity in the dorsal hippocampus. The present study used
in vivo MEMRI to map brain activation in control and rTg4510 mice under baseline, non-stimulated conditions. We observed significantly lower signal intensity, which may reflect underlying reductions in neuronal firing in rTg4510 versus control littermates. This effect appeared selective for memory formation structures, such as the hippocampus and amygdala, and was not observed in other ROI such as the septum, cerebellum, striatum, midbrain and brainstem. The hippocampus and amygdala play important roles in learning and memory, with the hippocampal structures contributing significantly to declarative memory while the amygdala contributes to implicit or emotional memory formation. While there is growing evidence that progressive reductions in hippocampal activity lead to generalized memory impairments, the implications for the amygdala are new for the rTg4510 mouse. The amygdala, particularly the basolateral regions, develop NFT neuropathology, however, a mechanism in rTg4510 mice remains to be thoroughly investigated. Dickey et al. (2009) immunohistochemically stained for tau and found no staining in thalamus. Indirect effects of tau pathology (or tau expression) in other regions, such as the neocortex, may have contributed to the differences observed here. In addition, we observed increases in signal intensity (activity) in the striatum, an effect that might reflect some changes in disinhibitory process arising from cortical inputs. However, despite the lack of any evidence of neuronal loss in the striatum, this region does accumulate NFTs beginning around 5.5 months [
26].
Progression of NFT pathology in rTg4510 mice has been considerably explored by several groups. Immunohistochemical analysis revealed that rTg4510 mice develop pre-tangle pathology at 2.5 M and fully formed NFTs at 4 M in the cortex and at 5.5 M in the hippocampus [
12,
28]. NFT pathology in the hippocampal formation occurs in a distinctly staged sub regional pattern beginning from CA1 to CA3 and dentate gyrus [
28]. Biochemical analysis of the rTg4510 hippocampus confirmed that the progression of NFT formation was drastically increased from 4 months to 6 months of age [
27,
29‐
31]. In agreement with these neuropathological observations, our
in vivo MRI analysis from rTg4510 mice (5.5-6.3 M) showed significant reduction of both neural activity and volume in the hippocampus. We observed significant reduction of signal intensity in the dentate gyrus of rTg4510 mice. Because CA1 pyramidal neurons were significantly decreased at this age, other sub regions of the hippocampus may receive less signaling associated with neural activity. Similar to our present observations, Yang et al. (2011) showed increased reactive astrocytic gliosis and enlarged microglia in CA1 from 8 M male rTg4510. One interesting possibility may be the sequestering of Mn
2+ in microglial cells occurs in rTg4510 CA1. This would either increase CA1 signal or pull away signal from the DG. An in-depth neurochemical investigation is warranted. An important follow up will be to control the expression of transgenic tau and the progression of NFT formation and/or neuronal loss through the administration of doxycycline and examine whether mice show unbalanced neural activity across hippocampal sub regions at different stages of disease progression.
A series of experiments using MEMRI have been carried out to examine axonal transport rates along the olfactory and optic tracts in transgenic mice for amyloid precursor protein (APP) expression. Using a fast temporal series of T
1 weighted image acquisitions it was shown that there is an age progressive decline in olfactory tract axonal transport rates of Mn
2+ that worsened after plaque formation. In a subsequent study it was demonstrated that loss of APP in knockout mice resulted in reduced axonal transport rates for Mn
2+, which was recovered by over-expressing human wildtype tau [
22]. Similar deficits in axonal transport rates have been reported in regions of the visual system and hippocampal formation of APP knockout mice [
24] and in the olfactory system of triple transgenic mice expressing APP and human presenilin [
23]. The above references illustrate the versatility of the MEMRI method. Localized
in vivo treatment can be used to track not only global brain activity as in the present work, but also transport rates on selective regions of the CNS [
32]. Our present results are consistent with recent work published by Kimura et al. (2007) showing lower activity in hippocampal and surrounding parahippocampal areas of aged mice with hyperphosphorylated tau protein. Levels of signal intensity in the parahippocampal areas highly correlated with performance on a spatial learning task [
33]. In the latter study Mn
2+ was provided to animals only 4 hours prior to imaging session and thus provides evidence of significant brain uptake of Mn
2+ across the blood brain barrier of mice. This is supported by
in vitro studies showing rapid uptake of Mn
2+ in brain, which appears to be aided by both active and passive transport processes [
34].
To our knowledge, this is the first MEMRI study examining neural activation in the rTg4510 mouse. Previous work on this transgenic mouse has focused on chemical and volumetric differences compared to controls. Both male and female mutant tau-expressing mice show reduced total brain volume which is largely associated with lower cortical and hippocampal volume [
25]. We observed this feature as well and also noted the previously reported widening of ventricles [
25]. This is consistent with postmortem tissue analysis in these same animals [
12,
26]. Proton MR spectroscopy showed greater myoinositol concentrations in the hippocampus and thalamus of rTg4510 mice, and this is consistent with increased gliosis [
25]. Interestingly, N-acetylaspartate, which is often taken as a marker of neuronal function, was not different between controls and rTg4510 mice. We observe here that
in vivo neural activity is reduced during baseline conditions in these animals. The previously reported lower number of neurons [
26], which may lead to lower hippocampal volume [
25], also may have accounted for the generally lower signal intensity in this region in rTg4510 mice. However, cortical neural activity was not different between control mice and rTg4510 mice despite the reduced volume in the latter. This could signify that the lower neural activity is due to reduced functional activity in memory structures. We also found that the reduction in hippocampal signal intensity is not uniform through the subanatomical layers of this paleocortical region. rTg4510 mice show a somewhat modest, but significant increase in signal intensity in the CA1 region. It is tempting to speculate on what such an effect may be associated with. One possibility could be due to compensatory changes in neuronal activity across the dorsal hippocampal areas; however, this would potentially result in maintenance of higher levels of performance on spatial navigation and memory tasks. Past data do not support such a contention at this time. One alternative might be linked to gliosis, which would be consistent with past work by Yang et al. (2012). Microglial cells migrating to this region might sequester Mn
2+ to such an extent that they result in increased signal intensity. This would not be due to increased neuronal activity, but rather it would be in line with the pathogenicity observed with tau over-expression at this age of the mice. This is somewhat supported by the literature. Quinolic acid lesions of neural tissue result in greater levels of Mn
2+ accumulation than in controls, thus supporting internalization of Mn
2+ in both neurons and glial cells [
35]. In conclusion, we find that mice over expressing human mutant tau(
P301L) show lower brain activity in selected regions involved in memory formation. The present novel findings are consistent with previous behavioral and biochemical studies in the rTg4510 mouse. Moreover, the methods used provide a novel approach to explore the specific relation between genetically-driven expression of pathogenic tau and functional changes over neurodegenerative disease progression.
Competing interests
The authors have no conflict of interest to disclose
Authors’ contributions
PDP, carried out the imaging data collection. GH, contributed to animal imaging and treatments. TK, contributed to data analysis and manuscript writing up. YR, carried out animal breeding, treatments and tissue processing. RB, carried out the immunohistochemistry, JL, contributed research animals manuscript editing and aided in conceptualization of study. MF, helped conceive the study, manuscript write up and did parts of the data analysis. NS, conceived study, manuscript write up and helped carry out portions of the study. All authors read and approved the final manuscript.