Alzheimer’s disease, described by the German neuropathologist Alois Alzheimer as ‘Dementia Praecox’ (Alzheimer, 1907) appears in both genetic (family AD) and sporadic forms. AD, the widespread cause of dementia, is characterised by progressive neurodegeneration and the appearance of specific histopathological markers represented by focal extracellular deposits of fibrillar β-amyloid (also called neuritic or senile plaques) in the brain parenchyma and in the wall of blood vessels, and by the intraneuronal accumulation of neurofibrillary tangles (NFT) formed as a result of the abnormal hyperphosphorylation of cytoskeletal Tau filaments (Selkoe, 2001). The initial neurodegenerative events of AD appear in the transentorhinal cortex, and subsequently spread to the entorhinal cortex and to the hippocampus. At the later stages of the disease the neurodegenerative process disseminates through the temporal, frontal and parietal lobes (Thompson et al. 2003, 2007). At these late stages the grey matter also undergoes severe atrophy manifested by a profound loss of neurons and synaptic contacts (Scheff et al. 2006).
Neurogenesis in post mortem tissues and in animal models of AD
Alzheimer’s disease, like every other form of dementia, is unique to humans; no other creature acquires AD (Toledano & Alvarez, 2004). Therefore, substantial efforts have been invested in generating relevant animal models of AD that reproduce various subsets of neuropathological, behavioural and/or biochemical alterations resembling those seen in human AD (Gotz et al. 2004; Cassel et al. 2008). The absence of a standard model makes it understandable, hence, that the data on neurogenesis arising from analysis of a post mortem AD brain (Haughey et al. 2002b; Jin et al. 2004b) as well as various transgenic mouse models of AD remain controversial.
Obviously, the changes in neurogenesis associated with AD are complex and require further clarification, as all the animal models studied exhibit only a reduced profile of AD pathology; thus, it is unclear whether the observed variations in neurogenesis are dependent on genotype in these studies (Wolf et al. 2006; Herring et al. 2009). Regarding the data from post mortem human tissues, these are intrinsically controversial and/or difficult to interpret because the post mortem material as a rule reflects the late stages of the disease. In addition, artefacts and misinterpretations can arise from post mortem delay, stage of the disease and treatments provided (these factors occurring either in isolation or combined). Furthermore, the discrepancies also depend on the methods used for labelling proliferating cells. For example, doublecortin labelling is known to identify both young and immature neurons (Chevallier et al. 2005; Verret et al. 2007). This alone may alter the interpretation of the data because more than 50% of the newborn cells die (Abrous et al. 2005). Labelling with 5-bromo-2′-deoxyuridine (BrdU) also suffers from uncertainties because it may not detect a distinct proliferative state, but instead mark repaired DNA in post-mitotic neurons and/or in cells entering an abortive cell cycle (Cooper-Kuhn & Kuhn, 2002; Rakic, 2002).
Neurogenesis in a triple transgenic mouse model of AD
Recently the triple transgenic mice model (3xTg-AD), which harbours three mutant genes for APPSwe, for PS1M146V and for tauP301L, was developed (Oddo et al. 2003a,b). These mice show temporal and region-specific Aβ and tau pathology, which closely resembles that seen in the human AD brain (Table 1). In addition, the 3xTg-AD animals show clear functional and cognitive impairments, including reduced LTP, as well as deficient spatial, long-term and contextual memory (Oddo et al. 2003a,b). The pathological changes start in the neocortex and expand towards the hippocampus (from 6 and 12 months old, respectively, for plaques, whereas NFTs start to be detectable from 12 months old; Rodríguez et al. 2008; Fig. 2). The functional deficits precede the appearance of histological hallmarks (Oddo et al. 2003a,b) and correlate with the accumulation of intraneuronal Aβ (Carroll et al. 2007; McKee et al. 2008; Fig. 2d,e). Therefore, the 3xTg-AD model is most accurate and relevant for studying the neurogenic and other pathological and physiological changes in AD, as even if there are other double transgenic animal models, such as the Tg2576xJNPL3 (APPSWE) mouse, and even if Tg2576 and VLW mice (Lewis et al. 2001; Ribe et al. 2005) reproduce plaques and NFTs (Table 1), they fail to completely mimic disease evolution by not developing early behavioural alterations and having ectopic (spinal cord) NFTs formation (Lewis et al. 2001), by not having plaques present throughout the whole hippocampus, and by not having long-lasting spatial memory deficits except at a very advanced (over 16 months) age (Ribe et al. 2005).
Table 1. Neuropathology in the main AD animal models.
|Lesion and transgenic mouse, rat and primate models||Neuropathology||Reference|
|Ageing||Cholinergic involution and amyloid deposition||Sani et al. (2003) Fischer et al. (1992) Michalek et al. (1989)|
|Electrolytic lesion||Neuronal death||Lescaudron and Stein (1999) Vale-Martinez et al. (2002)|
|Unspecific toxins (NMDA, ibotenic acid, quisalic acid, quinolic acid, colchicine, alkaloids, alcohol)||Neuronal death||Dunnett et al. (1991) Winkler et al. (1998) Boegman et al. (1985) Shaughnessy et al. (1994) Di Patre et al. (1989) Arendt (1994)|
|Specific toxins (AF64A, 192Ig-G saporin)||Cholinergic neuronal death||Waite et al. (1995) Chrobak et al. (1988) Hanin (1996) Wiley et al. (1991); Wiley (1992)|
|β-Amyloid||Cholinergic dysfunction||Giovannini et al. (2002) Pavia et al. (2000)|
|PS1M146L||Diffused plaques||Blanchard et al. (2003)|
|APP751SL||Plaques||Blanchard et al. (2003)|
|APP/Ld/2||Plaques||Moechars et al. (1996)|
|APPSwe||Plaques||Eriksen and Janus (2007)|
|APP Swedish, 695.K670N M671L||Plaques||Sturchler-Pierrat et al. (1997)|
|APP751SL/ PS1M146L||Plaques||Blanchard et al. (2003)|
|APPSWE/PS1dE9||Plaques||Savonenko et al. (2005)|
|APPSwedish and PS1M146L||Plaques||Janus et al. (2000)|
|APP695SWE||Plaques||Hsiao et al. (1996)|
|APPV717F||Plaques||Dodart et al. (2000)|
|K670N/M671L and V717F||Plaques||Janus et al. (2000)|
|APP Swedish, 695.K670N-M671L and Indiana V717F||Plaques||Eriksen and Janus (2007)|
|TgAPPsw and PS1 M146L||Plaques||Takeuchi et al. (2000)|
|APPSwedish and V717F||Plaques||Chishti et al. (2001)|
|V337M||Tangles||Tanemura et al. (2002)|
|4R/2N||Tangles||Tatebayashi et al. (2002)|
|TauP301L (4R,2-,3-)||Tangles||Lewis et al. (2000)|
|P301L||Tangles||Gotz et al. (2001)|
|TauP301L||Tangles||Arendash et al. (2004)|
|P301S/G272V||Tangles||Schindowski et al. (2006)|
|P301S||Tangles||Allen et al. (2002)|
|G272V, P301L, R406W||Tangles||Eriksen and Janus (2007)|
|Endogenous tau knocked out||Tangles||Andorfer et al. (2003)|
|P301L TET-off||Tangles||Ramsden et al. (2005)|
|7TauTg||Tangles||Ishihara et al. (2001)|
|Tg2576×JNPL3 (APPSWE)||Plaques and tangles||Lewis et al. (2001)|
|Tg2576 and VLW||Plaques and tangles||Ribe et al. (2005)|
|3xTg-AD||Plaques and tangles||Oddo et al. (2003b)|
|Tg478||None||Flood et al. (2009)|
|Tg1116||None||Flood et al. (2009)|
|K670M/N671L||Plaques||Kloskowska et al. (2010)|
|Tg478/Tg1116||Plaques||Flood et al. (2009)|
Figure 2. Brightfield micrographs showing the presence of β-amyloid within the pyramidal neurons of CA1 as well as the presence of a plaque in 12 months 3xTg-AD mice (b) compared with a non-Tg control animal (a). (c) We can see the accumulation phosphorylated Tau within the CA1 of 3xTg-AD mice. (d,e) Linear correlations between the mean number of cells containing β-amyloid in the hippocampal CA1 and the mean area density of HH3-positive cells in the GCL of the DG of male (d) and female (e) 3xTg-AD mice. S.Mol, stratum lacunosum molecular; S.Or, stratum oriens; S.Pyr, stratum pyramidale; S.Rad, stratum radiatum. Modified from Rodríguez et al. (2008) with permission.
Download figure to PowerPoint
In experiments on 3xTg-AD animals, we found impaired neurogenic capabilities in both the SVZ and SGZ of the hippocampal DG (Rodríguez et al. 2008, 2009). In both neurogenic niches, the SVZ and SGZ of the 3xTg-AD mice, a fair number of newly generated cells can be detected with different proliferation markers, such as BrdU, Ki67, PCNA and HH3 (Abrous et al. 2005; Rodríguez et al. 2008, 2009; Figs 3a–f and 4a,b,d). These newly-formed cells always show the distinctive characteristics of proliferating cells, being mainly localized in the inferior part of the GCL and demonstrating typical morphology such as irregular shape and small size (Figs 1b, 3a–f and 4a,b,d). Sometimes they tend to appear close together and/or form clusters (Abrous et al. 2005; Rodríguez et al. 2008, 2009), which rarely co-localize with glial fibrillary acidic protein (GFAP; the main glial cytoskeletal component; < 5%; Rodríguez et al. 2008, 2009; Figs 3c–f and 4d).
Figure 3. (a–d) Photomicrographs showing phosphorylated HH3 (a proliferating mitotic marker) within the DG of non-Tg mice. (a,b) Single labelling of HH3-positive cells (arrows) in the DG of 2 (a) and 12 months (b) non-Tg mice. (c,d) Dual labelling of HH3-positive cells (arrows) and glial cells (GFAP, blue) in the DG of 2 (c) and 12 months (d) non-Tg mice. Brightfield micrographs showing HH3-labelled cells within the DG of 3xTg-AD. (e,f) Dual labelling of HH3-positive cells (arrows) and glial cells (GFAP, red) in the DG of 2 (e) and 12 months (f) 3xTg-AD mice. (g,h) Bar graphs showing the mean area density HH3-labelled cells within the GCL of the dorsal DG of both 3xTg-AD and control non-Tg-AD mice males (g) and females (h). Asterisks indicate a significant difference in the means. GCL, granule cell layer; Mol.L, molecular layer. Modified from Rodríguez et al. (2008) with permission.
Download figure to PowerPoint
Figure 4. Brightfield micrographs showing phosphorylated HH3 (a proliferating mitotic marker) within the SVZ of both non-Tg and 3xTg-AD mice. (a,b) Single labelling of HH3-positive cells in the SVZ of 3 months non-Tg (a) and 3 months 3xTg-AD mice (b). (c) Bar graphs showing the mean area density of HH3-labelled cells within the SVZ of both control non-Tg and 3xTg-AD mice. Asterisks indicate a significant difference in the means. (d) Photomicrograph showing one of the few examples of co-localisation (marked by red circle) of HH3-positive cells (arrows) and glial cells (GFAP, blue/black; asterisks) in the SVZ of a 6 months non-Tg mouse, as the majority of proliferating cells did not have glial phenotype. CC, corpus callosum; CPU, caudate putamen nucleus; LV, lateral ventricles; SVZ, subventricular zone. Modified from Rodríguez et al. (2009) with permission.
Download figure to PowerPoint
In the GCL of the hippocampal DG of 3xTg-AD mice (as in normal aging; Kuhn et al. 1997; Abrous et al. 2005), the rate of neurogenesis starts to decrease from the age of 6 months (over 50% reduction when compared with early ages such as 2 months) and decreases further at later ages (Fig. 3g,h), affecting more females than males (Rodríguez et al. 2008). The age-dependent decrease in neurogenesis, however, is much more prominent in the 3xTg-AD mice when compared with the controls. From 9 to 12 months old, both genders retained very little capacity of forming new cells within the GCL; in non-Tg control animals the neurogenic levels that account approximately for a 20–35% of the young age levels were still preserved (Fig. 3g,h). A significant decrease in neurogenesis was also observed in the SVZ during normal aging, but at a lower level ranging between 19% and 31% (Rodríguez et al. 2009; Fig. 4a–c). The 3xTg-AD when compared with normal animals presented a further 40% decrease (Fig. 4c), that appears as early as 3 months old, and is sustained through later ages (Rodríguez et al. 2009).
Studies in the 3xTg-AD model also demonstrated that female mice are affected earlier than males (4 vs. 9 months old; Fig. 3g,h; Rodríguez et al. 2008). These findings are not only in line with the recently reported sexual dimorphism observed in cognitive performances (Clinton et al. 2007), but also correlate to the well-known fact that AD affects women earlier and with more severity than men (Baum, 2005; Webber et al. 2005). Several lines of evidence suggest that this difference, even with a potential similar disruption mechanism, might be exacerbated by the circulating levels of oestrogens (Manly et al. 2000; Baum, 2005; Webber et al. 2005) as a result of the endocrine status (Galea et al. 2006).