sAPPα Rescues Age-Linked Decline in Neural Progenitor Cell Proliferation

Introduction

Amyloid precursor protein (APP) can undergo metabolism through two distinct pathways leading to the production of a number of intra- and extracellular metabolites. The amyloidogenic pathway of APP cleavage, so termed for its involvement in the formation of β-amyloid (Aβ), begins with β-secretase, beta-site APP cleaving enzyme 1 (BACE1), and releases soluble amyloid precursor protein β (sAPPβ) to the lumen [for review (Vassar, 2012)]. In the non-amyloidogenic APP cleavage pathway, initial cleavage occurs through α-secretase enzymes, primarily composed of a disintegrin and matrix-metalloproteinase proteins (ADAM), leading to the release of sAPPα (Buxbaum et al., 1998; Lammich et al., 1999; Asai et al., 2003; Jorissen et al., 2010). It has been previously reported that sAPPα has trophic properties in a number of different cell types including fibroblasts (Saitoh et al., 1989), thyroid epithelial cells (Pietrzik et al., 1998), embryonic stem cells (Ohsawa et al., 1999) and carcinoma cells (Ko et al., 2004). With respect to adult neural progenitor cells (NPC), sAPPα has been shown to bind to and stimulate the proliferation of NPC both in vitro and in vivo (Caille et al., 2004; Demars et al., 2011).

In the aging brain, there is a marked decline in neurogenesis in both the subventricular zone (SVZ) (Mirich et al., 2002; Shook et al., 2012) and subgranular layer (SGL) (Kuhn et al., 1996; Cameron and McKay, 1999; Bernal and Peterson, 2004; Bondolfi et al., 2004; Kronenberg et al., 2006; Ben Abdallah et al., 2010; Encinas et al., 2011; Miranda et al., 2012). This decline in neurogenesis has been shown to manifest in deficits in olfactory function and hippocampaldependent learning and memory (Bizon et al., 2004; Enwere et al., 2004; Dupret et al., 2008). However, the cause of neurogenic decline in the aging brain remains somewhat controversial. Some studies argue that the proliferation of NPC and their maturation declines with age (Kuhn et al., 1996; Heine et al., 2004; Rao et al., 2005; Morgenstern et al., 2008). Olariu and colleagues suggest that this decline is not due to alterations in the length of the cell cycle of NPC in the SGL (Olariu et al., 2007). Encinas and colleagues suggest that neurogenic decline is due to the disappearance of NSC by their conversion into mature hippocampal astrocytes (Encinas et al., 2011). Bonaguidi et al. (2008) suggest that upregulation of bone morphogenetic protein (BMP) signaling or downregulation of its antagonists may underlie age-linked reduced self-renewal of NSC leading to the decline in neurogenesis (Bonaguidi et al., 2008). Other groups have argued that the rate of proliferation does not decline with aging, but the cells may exhibit increased quiescence that may be attributed to a decline in the vascular niche (Hattiangady and Shetty, 2008). Although this debate is still ongoing, it is apparent that the neurogenic niche undergoes changes associated with aging that have the potential to underlie neurogenic decline (Ahlenius et al., 2009; Villeda et al., 2011; Miranda et al., 2012). For instance, it has been well established that the expression of many growth factors and their receptors that are integral to neurogenic processes peak during development and decline in expression thereafter [for review (Klempin and Kempermann, 2007)]. Neurogenesis in the adult central nervous system (CNS) plays a role in hippocampus- and olfaction-dependent learning and memory [for review (Kempermann and Gage, 2000; Bernal and Peterson, 2004; Lie et al., 2004)]. These findings raise the possibility that reduced neurogenesis may, at least in part, account for impaired learning and memory and cognitive deterioration in the elderly (Seki and Arai, 1995; Kuhn et al., 1996; Tropepe et al., 1997; Kempermann et al., 1998; Kempermann et al., 2002) and may enhance vulnerability to Alzheimer’s disease [(AD) for review see (Lazarov et al., 2010)].

Aging remains the predominant risk factor for the development of sporadic forms of AD. Alterations in the cleavage pattern of APP are causative of familial forms of the disease [for review (Selkoe, 2001)]. We have previously reported that deficits in proliferation and neurogenesis occur long before the appearance of pathological hallmarks or cognitive impairments in a mouse model of familial Alzheimer’s disease (Demars et al., 2010). Taken together with our observation that sAPPα is a proliferation factor of adult NPC, this may suggest the hypothesis that increasing levels of sAPPα in the neurogenic niches would rescue neurogenic deficits.

Here we show that a dramatic decline in neurogenesis takes place at the age of 7–9 months, compared to mice at 2 months of age, and only a minor further decrease takes place at age of 20 months. This suggests that the significant decline in neurogenesis takes place during adulthood, rather than during aging. This decline is characterized by reduced numbers of fast proliferating NPC (type II cells in the SGL and type C cells in the SVZ), but no decline in the number of NSC. Intracerebroventricular (ICV) injection of sAPPα ameliorates proliferation deficits in adult mice. In support of that, adding sAPPα to a neurosphere culture derived from 7–9 month old or 20 month old mice enhances NPC proliferation. Additionally, we show that at similar concentration range, sAPPβ does not evoke the same proliferative effect as sAPPα. This study suggests that therapy aimed at rescuing decline in neurogenesis should be applied during adulthood rather than aging, and that enhancement of sAPPα levels may rescue this decline.

Materials and methods

Results

To examine the course of decline in neurogenesis with age we first assessed the extent of neurogenesis in the SGL and SVZ of 2 (young), 7–9 (adult) and 20 (aging) months old mice (N=4). For this purpose, we quantified the number of type B (nestin+GFAP+), total fast proliferating cells (BrdU+), type C (BrdU+nestin+, BrdU+DCX+), and type A (BrdU-DCX+) cells. Unbiased stereology revealed a severe reduction in the number of total proliferating cells (Figure 1A), and the number of type C and A cells by 7–9 months of age in the SVZ (Figure 1B–D). Notably, the number of fast proliferating cells (BrdU+) and NPC (nestin+BrdU+) is significantly reduced by 7–9 months of age, and stays reduced without further significant reduction by 20 months of age (Figure 1A,B). This may suggest that the reduced proliferation of NPC in the SVZ by 7–9 months of age might play a major role in reduced extent of neurogenesis. The number of neuroblasts and immature neurons is significantly reduced by 7–9 months, and this number is further reduced by 20 months (Figure 1C,D,Kiii–v). The number of type B cells (GFAP+nestin+) was not significantly changed in any of the age groups (Figure 1E), suggesting that the pool of NSC in the SVZ does not decline with age.

Examination of the cell population in the SGL of the different ages revealed similar observations. We quantified the number of type I (nestin+GFAP+), total fast proliferating cells (BrdU+), type II (BrdU+nestin+, BrdU+DCX+), and type III (BrdU-DCX+) cells. The number of total proliferating cells (Figure 1F), and the number of type II and III cells (Figure 1G–I) was dramatically reduced by 7–9 months of age in the SGL. The number of type II and III cells remains reduced. The number of proliferating neuroblasts, BrdU+DCX+, is reduced severely by 7–9 months of age, with a further significant reduction in their number between middle and old age (Figure 1H,Kiii,iv). The same is true for the total number of post-mitotic immature neurons (BrdU-DCX+). A dramatic reduction in this population occurs between 2 and 7–9 months of age with a smaller but significant reduction taking place thereafter up to 20 months of age (Figure 1I). Of note, the number of total proliferating BrdU+ cells and of fast proliferating NPC type II (BrdU+nestin+, BrdU+DCX+) was dramatically reduced. Notably, there is no change in the number of NSC (Figure 1J), suggesting that the pool of NSC in the SGL does not decline with age. We and others have shown that sAPPα is a proliferation factor of NPC in the adult brain (Caille et al., 2004; Baratchi et al., 2011; Demars et al., 2011). Thus, in light of the dramatic reduction in proliferation of NPC in the SVZ and the SGL of mice at age of 7–9 months, we asked whether addition of sAPPα into the neurogenic areas can ameliorate deficits in proliferative cells in the brains of 7–9 months old mice in vivo. To this end, we performed intracerebroventricular (ICV) injection of recombinant sAPPα or PBS (supplementary Figure 1) in mice at 7–9 months of age followed 6 hours later with a single i.p. dose of BrdU and sacrifice 24 hours later (N=4). Single ICV injection of sAPPα in 7–9 month old mice dramatically increased the number of proliferating (BrdU+) cells in both the SVZ (Figure 2A,Ii,ii) and SGL (Figure 2E,Iiii,iv). To assess the impact of ICV sAPPα injection on NPC proliferation, we quantified the number of cells co-labeled with BrdU and nestin. We show that sAPPα injection rapidly and significantly increased the number of proliferating NPC in both the SVZ (Figure 2B,Ivi) and SGL (Figure 2F) of 7–9 month old mice. Notably, in the SVZ, the number of proliferating neuroblasts was increased following sAPPα injection (Figure 2C,Iv), suggesting that enhanced proliferation had a direct effect of the number of neuronaly- committed NPC. In the SGL, the number of neuroblasts appeared albeit reduced but not significantly changed (Figure 2G). Interestingly, there was a statistically insignificant reduction in the number of NSC (GFAP+/Nestin+) in the SVZ (Figure 2D). In the SGL, reduction in the number of NSC was significant (Figure 2H). This may raise the possibility that sAPPα induces NSC asymetric division resulting in more NPC in vivo. Alternatively, sAPPα may facilitate their commitment to neuronal lineage. Taken together, these results suggest that sAPPα can enhance aging-linked reduced proliferation in both the SVZ and SGL through stimulating NPC proliferation.

To further establish that the addition of sAPPα stimulates NPC proliferation, we isolated NPC from the SVZ of mice at age of 2, 7–9 or 20 months old, and treated the neurosphere cultures with either recombinant sAPPα or vehicle. Consistent with previous reports suggesting proliferation deficits could be masked by culture conditions (Ahlenius et al., 2009; Bouab et al., 2011), we saw no significant difference in neurosphere diameter or total NPC number in neurosphere cultures derived from mice that are 2, 7–9 or 20 months old (Figure 3). However, we could stimulate proliferation with the addition of 10nM sAPPα in neurospheres derived from 2, 7–9 month and even 20 months old mice as indicated by increased neurosphere diameter (Figure 3A,C) and increased total NPC numbers (Figure 3B). This result suggests that NPC derived from mice at various ages retain the ability to proliferate in response to sAPPα in vitro.

The cleavage of APP can produce either sAPPα or sAPPβ. Therefore, we sought to determine if NPC respond differentially to the two sAPP species. We assayed sphere diameter of neurospheres derived from single-dissociated NPC treated with varying concentrations of either sAPPα or sAPPβ. Treatment with sAPPα increased neurosphere diameter at 10nM concentrations with a trend toward significance at 1nM. However, sAPPβ treatment failed to significantly increase proliferation at any of the concentrations tested (Figure 4A, representative images C). In our previous work, we showed that sAPPα is able to ameliorate deficits in NPC proliferation caused by matrix-metalloproteinase inhibition that inhibits enzymes exhibiting α-secretase activity that yields sAPPα production (Demars et al., 2011). The use of the inhibitor is meant to inhibit endogenous production of sAPPα, thus preventing a masking of the proliferative effect exerted by recombinant sAPPα. Therefore, we sought to determine whether sAPPβ would enhance proliferation in the absence of endogenous production of sAPPα. As previously reported, 1 µM GM6001 MMP inhibitor treatment impairs proliferation, while sAPPα ameliorates these proliferation deficits [Figure 4B, (Demars et al., 2011)]. At 1 nM concentrations, sAPPβ enhances proliferation to the same extent as sAPPα. However, at concentrations above 10 nM, sAPPβ treatment has no effect on rescue of MMP inhibitor induced proliferation deficits (Figure 4B).

Treatment with MMP inhibitor likely exclusively reduces sAPPα. Thus, we next asked whether endogenous production of sAPPβ masks the proliferative effect of recombinant sAPPβ. For this purpose, we inhibited endogenous production of sAPPβ by treating NPC culture with β-secretase inhibitor IV. We show that treatment of NPC with β-secretase inhibitor reduces proliferation dramatically as assayed by neurosphere diameter and total cell number (Figure 5A). While sAPPα ameliorates this deficit at 10 nM concentrations, 1 nM sAPPβ, the most effective dose in previous experiments, does not enhance proliferation following β-secretase inhibition (Figure 5A, B). To further exclude the possibility that the presence of either endogenous sAPPα or sAPPβ mask the proliferative effect of recombinant sAPPβ, NPC cultures were treated with both MMP and β-secretase inhibitor. Treatment of NPC with GM6001 or β-secretase inhibitor IV reduced neurosphere diameter (Figure 6A) and total cell number (Figure 6B) to the same extent following 7DIV. To assess the ability of sAPPα or sAPPβ to stimulate proliferation under conditions of minimal endogenous sAPP, we added exogenous recombinant sAPPα or sAPPβ to NPC under dual secretase inhibition. sAPPα rescued proliferation deficits to control levels. Conversely, sAPPβ failed to ameliorate the proliferation deficits incurred from dual secretase inhibition (Figure 6A,B). These results suggest that sAPPα is a more potent proliferation factor for NPC than sAPPβ.

Finally, we sought to address whether, similarly to sAPPα, sAPPβ would rescue age-dependent proliferation deficits of NPC. Because age-dependent proliferation deficits are not apparent in vitro, we examined the effect of sAPPβ administration on proliferation in 7–9 month old mice compared to the young in vivo (N=4). We show that ICV injection of sAPPβ led to a markedly reduced population of proliferating cells in both the SVZ (Figure 7A–C) and SGL (Figure 7D–F), including a dramatically reduced population of BrdU+/DCX+ proliferating neuroblasts (Figure 7B SVZ, Figure 7E SGL). Interestingly, similarly to the sAPPα injected cohorts, sAPPβ injected animals showed a decrease in GFAP+/nestin+ NSC in the SGL (Figure 7F) and a trend in the SVZ, albeit not significant (Figure 7C). These results indicate that at equimolar concentrations, sAPPα and sAPPβ have significantly different effects on proliferating cells in the two neurogenic niches. While sAPPα is able to ameliorate deficits in proliferation, sAPPβ reduces the number of proliferating cells in the already depleted brains of 7–9 month old animals.

Discussion

This study reports several important observations. First, there is a dramatic decline in neurogenesis in adulthood, prior to aging. This is consistent with previous reports on neurogenesis in the aging brain (Cameron and McKay, 1999; Bernal and Peterson, 2004; Kronenberg et al., 2006). Second, the number of proliferating cells is dramatically reduced by 7 months of age in both the SVZ and SGL, with no further significant decline. We further show that the decline in proliferation is due, in large part, to a reduction in the number of proliferating NPC without any change in the number of NSC in either region. The molecular mechanism underlying this decline in proliferation is not fully elucidated. It is possible that the level of essential proliferation factors is declined with age. Notably, binding sites for sAPP are localized to rapidly proliferating-C cells in the adult brain (Caille et al., 2004). However, our results cannot rule out the possibility that the age-linked decline in proliferating cells is due to increased quiescence of NSC and a reduced NPC pool. As NSC proliferate very slowly (Zheng et al., 2004), our paradigm of single-pulse BrdU injection likely will not capture this population. However, our results do indicate that the total number of NSC is steady across all ages and suggest that a waning number of NSC is not the cause of neurogenic decline.

Third, we show that a single ICV injection of recombinant sAPPα is sufficient to significantly ameliorate age-linked deficits in the number of proliferative NPC. It should be noted that while the final concentration of sAPP following injection is uncertain, it was designed to be diluted approximately 100 fold in the cerebrospinal fluid of the ventricular space. Given the injected dose of 1 µM, this would result in a final concentration of approximately 10 nM, the most effective concentration of sAPPα in our culture conditions. The injection of sAPPα significantly increased the number of BrdU/nestin co-labeled NPC in both the SGL and SVZ. A single injection of sAPPα into the lateral ventricle enhanced the number of neuroblasts in the SVZ but not in the SGL. One possibility is that because of the relatively low number of neuroblasts in the SGL compared to the SVZ, a single injection is not sufficient to evoke an increase in neuroblasts within 24 hours. It is reasonable to assume that some of the detected BrdU+DCX+ started to proliferate prior to the stimulation of recombinant sAPPα. Thus, more studies are warranted in order to determine whether injection of sAPPα leads to increased number of new neurons and to their survival and functional incorporation in the olfactory bulb and granular layer of the dentate gyrus.

Fourth, we show that sAPPβ is a less potent proliferation factor than sAPPα in vitro and may only function to stimulate proliferation in a relatively small concentration range. Several of the mutations in APP, that are causative of FAD, lie in close proximity to the β-secretase cleavage site at the N-terminal portion of the Aβ region. These mutations cause a shift in the metabolism of APP toward the amyloidogenic pathway, increasing sAPPβ production at the expense of sAPPα (Thinakaran et al., 1996). FAD-linked transgenic mice display impaired proliferation of NPC prior to the onset of pathological hallmarks or the presentation of memory deficits. In vitro, NPC derived from these mice have impaired proliferation, suggesting a potential intrinsic mechanism caused by the mutations (Demars et al., 2010). The differential activity between sAPPα and sAPPβ has been previously reported with respect to other trophic properties such as neuroprotection against glutamatergic or β-amyloid toxicity (Furukawa et al., 1996), promotion of axonal elongation and primary dendritic length (Chasseigneaux et al., 2011), hippocampal long-term potentiation (LTP) (Taylor et al., 2008) and rescue of prenatal lethality in an APP/APLP2 knockout mouse (Li et al., 2010; Weyer et al., 2011). However, the cause of this functional divergence between two peptides that differ by only 16 amino acids has yet to be elucidated. The C-terminal portion of sAPPα alone is unable to stimulate proliferation in embryonic stem cells (Ohsawa et al., 1999). Whatever the cause of these differences, it is feasible that a shift in the homeostatic balance of APP processing favoring the amyloidogenic pathway would impair neurogenesis. Not only would this result in the production of the less trophic sAPPβ but it would also lead to increases in the transcriptionally active form of APP intracellular domain (AICD), which has been shown to be preferentially produced by this pathway (Goodger et al., 2009; Belyaev et al., 2010) and to be a negative regulator of proliferation (Ma et al., 2008; Ghosal et al., 2010). Therefore a shift in the balance of these cleavage pathways could underlie, at least in part, neurogenic deficits in aging and AD.

Finally, we show that ICV injection of sAPPβ at similar/equimolar concentrations to sAPPα does not enhance the number of proliferating cells in either the SVZ or SGL of the aging brain. Conversely, sAPPβ injection reduced the number of proliferating cells. There is evidence from peripheral neurons that during times of reduced growth factor support, sAPPβ is released and binds to DR6 inducing neurodegeneration (Nikolaev et al., 2009). The mechanism underlying this suppression of proliferation is not known. Future experiments will be aimed at unraveling the significance of sAPPβ regulated decline in proliferation. It should be noted that the concentration of sAPPβ used was designed to be an equimolar dose to sAPPα injections. As we have shown in our in vitro assays the sAPPs may have different optimal concentrations thus we may have not captured the optimal dose in our experiments. Alternatively, a difference in the half-life of sAPPα versus sAPPβ or their rate of metabolism may underlie this outcome.

In summary, this study shows that extent of neurogenesis declines well prior to aging, that the decline in neurogenesis as a function of age is largely due to a decline in the number of proliferating NPC and that this reduction can be reversed by a single-dose ICV injection of sAPPα, while the same dose of sAPPβ resulted in further impairment of proliferation. These findings highlight the differential activity of the sAPPs with respect to proliferation of NPC. Together with previous studies, this work provides evidence that the regulation of APP processing plays a major role in the regulation of NPC proliferation in the adult brain. This paves the way for therapeutic intervention with an emphasis on maintaining a homeostatic balance in APP processing both in AD and physiological aging. Though aging is the predominant risk factor for sporadic forms of AD, little is known about the expression of APP and APP metabolites during normal aging. One study of note did examine APP maturation and processing during cellular aging in a human lung fibroblast cell line and showed a decreasing metabolism with increasing cellular age including a decline in sAPPα (Kern et al., 2006). This group further showed that increasing membrane cholesterol levels correlated with increasing cellular age. Intriguingly, APP, BACE1 and presenilin1 have been shown to associate increasingly with detergent resistant, cholesterol rich, membranes or “lipid rafts” with increasing cellular age (Kang et al., 2006). Evidence suggests that amyloidogenic processing of APP is sequestered in these detergent resistant membrane micro-domains (Simons et al., 1998; Schneider et al., 2008). Thus the possibility exists that in both physiological aging and AD there is a shift in the metabolic pathway or cleavage pattern of APP favoring the amyloidogenic pathway over the seemingly more trophic non-amyloidogenic pathway.

Supplementary Material